An RNA-Centric Global View of Clostridioides Difficile Reveals Broad 2 Activity of Hfq in a Clinically Important Gram-Positive Bacterium

bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

1 An RNA-centric global view of Clostridioides difficile reveals broad 2 activity of Hfq in a clinically important Gram-positive bacterium

3 Manuela Fuchs 1§, Vanessa Lamm-Schmidt 1§, Falk Ponath 2, Laura Jenniches 2, Lars Barquist 2,3, 4 Jörg Vogel 1,2, Franziska Faber 1, *

5

6 1 Institute for Molecular Infection Biology (IMIB), Faculty of Medicine, University of Würzburg, D- 7 97080 Würzburg, Germany

8 2 Helmholtz Institute for RNA-based Infection Research (HIRI), D-97080 Würzburg, Germany

9 3 Faculty of Medicine, University of Würzburg, D-97080 Würzburg, Germany

10 § authors contributed equally

11 * Correspondence: 12 Franziska Faber: phone +49-931-3186280; email: [email protected]

13

14 Running title: primary transcriptome of Clostridioides difficile 15 Keywords: Clostridioides difficile, Transcriptional start site, Transcriptional termination site, 16 differential RNA-seq, noncoding RNA, antisense RNA, small RNA, post-transcriptional control, 17 Hfq bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

18 ABSTRACT

19 The Gram-positive human pathogen Clostridioides difficile has emerged as the leading cause of 20 antibiotic-associated diarrhea. Despite growing evidence for a role of Hfq in RNA-based gene 21 regulation in C. difficile, little is known about the bacterium’s transcriptome architecture and 22 mechanisms of post-transcriptional control. Here, we have applied a suite of RNA-centric 23 techniques, including transcription start site mapping, transcription termination mapping and 24 Hfq RIP-seq, to generate a single-nucleotide resolution RNA map of C. difficile 630. Our 25 transcriptome annotation provides information about 5’ and 3’ untranslated regions, operon 26 structures and non-coding regulators, including 42 sRNAs. These transcriptome data are 27 accessible via an open-access browser called ‘Clost-Base’. Our results indicate functionality of 28 many conserved riboswitches and predict novel cis-regulatory elements upstream of MDR-type 29 ABC transporters and transcriptional regulators. Recent studies have revealed a role of sRNA- 30 based regulation in several Gram-positive bacteria but their involvement with the RNA-binding 31 protein Hfq remains controversial. Here, sequencing the RNA ligands of Hfq reveals in vivo 32 association of many sRNAs along with hundreds of potential target mRNAs in C. difficile providing 33 evidence for a global role of Hfq in post-transcriptional regulation in a Gram-positive bacterium. 34 Through integration of Hfq-bound transcripts and computational approaches we predict 35 regulated target mRNAs for the novel sRNA AtcS encoding several adhesins and the conserved 36 oligopeptide transporter oppB that influences sporulation initiation in C. difficile. Overall, these 37 findings provide a potential mechanistic explanation for increased biofilm formation and 38 sporulation in an hfq deletion strain and lay the foundation for understanding clostridial ribo- 39 regulation with implications for the infection process.

2 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

40 INTRODUCTION

41 Antibiotic-resistant bacteria are one of the major global threats to human health, endangering our 42 ability to perform a range of modern medical interventions. The obligate anaerobe, spore-forming 43 Clostridioides difficile (C. difficile) has become the leading cause of antibiotic-associated diarrhea 44 over the past two decades (Rupnik et al., 2009). In addition, C. difficile shows increasing numbers 45 of multi-resistant clinical isolates (Peng et al., 2017) and recurrent infections after antibiotic 46 therapy, which often leaves fecal microbiota transfer as the only clinical option (Khanna and 47 Gerding, 2019).

48 The clinical challenges posed by C. difficile have prompted much effort to understand how 49 this pathogen regulates virulence in response to environmental conditions. As a result, there is a 50 comprehensive body of literature focusing on toxin production and sporulation control by several 51 global metabolic regulators including CcpA, CodY, Rex and PrdR (Bouillaut et al., 2015; Martin- 52 Verstraete et al., 2016). Furthermore, several specialized and general sigma factors including 53 TcdR (Martin-Verstraete et al., 2016), Spo0A (Pettit et al., 2014), SigD (El Meouche et al., 2013; 54 McKee et al., 2013), SigH (Saujet et al., 2011) and SigB (Kint et al., 2017) have been linked to 55 virulence and metabolism, although the exact molecular mechanisms often remain unknown. 56 Importantly, most of this knowledge has been accumulated through detailed studies of individual 57 genes and promoters. By contrast, RNA-seq based annotations of the global transcriptome 58 architecture, which have accelerated research in the Gram-positive pathogens Listeria (Wurtzel 59 et al., 2012), Staphylococcus (Mader et al., 2016) and Streptococcus (Warrier et al., 2018), have not 60 been available for C. difficile so far.

61 This paucity of global knowledge about RNA output in C. difficile readily extends to post- 62 transcriptional control of gene expression. The bacterium is of particular scientific interest, being 63 the only Gram-positive species thus far in which deletion of hfq seems to have a large impact on 64 gene expression and bacterial physiology (Boudry et al., 2014; Caillet et al., 2014). Specifically, 65 deletion of hfq increases sporulation (Maikova et al., 2019), a crucial pathogenic feature of this 66 bacterium that enables transmission between hosts. In Gram-negative bacteria, Hfq commonly 67 exerts global post-transcriptional control by facilitating short base pairing interactions of small 68 regulatory RNAs (sRNAs) with trans-encoded target mRNAs (Holmqvist and Vogel, 2018; Kavita 69 et al., 2018) but its role in Gram-positive bacteria remains somewhat controversial. For example, 70 recent work on the function of Hfq in Bacillus subtilis (B. subtilis), the model bacterium for 71 Firmicutes, revealed in vivo association with a subset of sRNA (Dambach et al., 2013), but 72 comparative analyses of a wild-type and hfq knockout strain revealed only moderate effects on 73 sRNA and mRNA transcript levels (Hammerle et al., 2014) and the absence of any significant

3 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

74 growth defect (Rochat et al., 2015) which lead to the conclusion that Hfq plays a minor role in 75 post-transcriptional regulation in B. subtilis.

76 Previous efforts using in silico methods (Chen et al., 2011) and RNA-sequencing 77 (Soutourina et al., 2013) have predicted >100 sRNA candidates in C. difficile, which would suggest 78 the existence of a large post-transcriptional network. However, under which conditions these 79 sRNAs are expressed, which targets they regulate, and whether they depend on Hfq remain 80 fundamental open questions.

81 Another important feature of post-transcriptional control in C. difficile are cis-regulatory 82 RNA elements. There has been pioneering work deciphering the function of genetic switches 83 located in the 5’ untranslated region (5’ UTR) of the flgB operon, containing the early stage 84 flagellar genes, and the cell wall protein encoding gene cwpV (Anjuwon-Foster and Tamayo, 2017; 85 Emerson et al., 2009). Moreover, cyclic di-GMP responsive riboswitches were shown to regulate 86 biofilm formation and toxin production (Bordeleau et al., 2011; McKee et al., 2018; McKee et al., 87 2013; Peltier et al., 2015). However, other members of the many different riboswitch classes or 88 common cis-regulatory elements such as RNA thermometers have not been systematically 89 searched for. Combined with the nascent stage of sRNA biology in C. difficile, this argues that global 90 approaches are needed to understand the full scope of post-transcriptional regulation in this 91 important human pathogen.

92 In the present study, we have applied recently developed methods of bacterial RNA 93 biology (Hor et al., 2018) to construct a global atlas of transcriptional and post-transcriptional 94 control in C. difficile. Our approach demonstrates how integration of different RNA-sequencing 95 based methods readily reveals new regulatory functions. Accordingly, we were able to predict 96 regulatory interactions between a newly annotated Hfq-binding sRNA and target mRNA 97 candidates. Overall, we provide evidence for extensive Hfq-dependent post-transcriptional 98 regulation and provide the foundation for future mechanistic studies of RNA-based gene 99 regulation in C. difficile.

100 RESULTS

101 High-resolution transcriptome maps of C. difficile 630

102 For generating high-resolution transcriptome maps of C. difficile, we chose the toxigenic reference 103 strain 630 (DSM 27543, GenBank: CP010905.2). Being widely used by the C. difficile community, 104 this strain offers the most comprehensive genome annotation. Using two different global RNA-seq 105 approaches, we analyzed RNA samples from three different conditions: late-exponential and 106 early-stationary phase of growth in Tryptone-yeast broth, as well as late-exponential phase in

4 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

107 Brain-heart-infusion broth (Fig. 1A). The resulting genome-wide maps provide single-nucleotide 108 resolution transcriptional start sites (TSSs), transcript ends, 5’ and 3’ untranslated regions (UTRs) 109 and operon structures (Fig. 1B). In addition, we have used this information to correct previous 110 ORF annotations, to add previously overlooked small genes and to annotate sRNA loci. Inspired 111 by other online gene expression databases such as SalCom (Kroger et al., 2013), AcinetoCom 112 (Kröger et al., 2018) and Theta-Base (Ryan et al., 2020), we have launched an open-access 113 interactive web browser, called ‘Clost-Base’ (https://www.helmholtz- 114 hiri.de/en/datasets/clostridium), for our transcriptome data. This online resource for C. difficile 115 allows easy visualization of the transcriptomic data in the context of annotated coding and non- 116 coding genes as well as transcript features (e.g. transcription start sites and transcription 117 termination sites) that we have experimentally determined in this work. The browser enables 118 search queries and retrieval of primary sequences for any annotated feature in the database and 119 is a valuable resource for the C. difficile community.

120 Genome-wide annotation of transcription start sites

121 Differential RNA-seq (dRNA-seq) (Sharma et al., 2010; Sharma and Vogel, 2014) was performed 122 to capture 5’ ends of transcripts, a method relying on the differential treatment of input RNA 123 sample with terminator exonuclease (Tex). In brief, one half of the sample remains untreated 124 (Tex- - -

125 other) half to capture is treated both (Tex+) primary leading (5′ toPPP) the and specific processed degradation (5′ P or of 5′OH) processed 5’ ends of- transcripts; the 126 thus enriching primary transcripts and enabling TTS annotation (Fig. 1C). (5′Conversely,P or 5′OH) relative RNAs, 127 read enrichment in the Tex- cDNA libraries indicate RNA processing sites (Fig. 1C). With a genome 128 size of 4,274,782 bp, the C. difficile 630 genome comprises annotations for 3,778 predicted coding 129 sequences as well as tRNAs, rRNAs and the housekeeping RNAs 6S RNA, RNaseP, tmRNA and SRP. 130 By identifying 2,293 transcriptional start sites (TSS), we were able to define transcriptional units 131 of individual genes and polycistronic operons for approximately half of the genome (Table S1).

132 Benchmarking our global data with previous studies of individual genes (Emerson et al., 133 2009; Saujet et al., 2011; Wydau-Dematteis et al., 2018), our dRNA-seq results are fully consistent 134 with published TSSs of sigH, cwpV, cwp19, and spo0A (Fig. 1D). For annotation purposes, we 135 assigned TSS to one of five classes according to their genomic location and expression level: pTSS 136 (primary TSS of a gene or operon), sTSS (secondary TSS showing lower expression level compared 137 to pTSS for the same gene or operon), iTSS (internal TSS located inside a gene), aTSS (antisense 138 TSS to a gene within 100 nt distance) and oTSS (orphan TSS, no nearby gene) (Fig. 2A). Naturally, 139 some of these TSS annotations overlap, for example, among the 1627 pTSS, 126 are located within 140 a gene (iTSS) and 46 are transcribed antisense (aTSS) to a gene. However, in contrast to other

5 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

141 bacteria where antisense transcription is a pervasive transcriptome feature (Wade and Grainger, 142 2014), it only accounts for approx. 6% of TSS events in C. difficile.

143 Internal TSS were abundant and accounted for 20% of all start sites, many of which seem 144 to uncouple downstream genes within operons. However, they may also indicate mis-annotated 145 translational start codons. One example is a putative hydrolase (CDIF630_02227) encoded 146 between ermB1 and ermB2 of the erythromycin resistance cassette. We detected an iTSS located 147 14 bases downstream of the presently annotated canonical start codon. However, we manually 148 checked for canonical (AUG) or alternative start codons (UUG, GUG) downstream of the annotated 149 start codon, including a ribosome-binding site (RBS). This approach resulted in the re-annotation 150 of the CDIF630_02227 ORF with the new canonical start codon 43 bases downstream of the 151 existing ORF annotation, placing a RBS in optimal distance to the new start codon and generating 152 a 5’ UTR length of 29 nucleotides (Fig. 2B). Similarly, we propose re-annotations in eight 153 additional cases, including the spaFEGRK operon encoding an antibiotic/multidrug-family ABC 154 transport system (Table S2).

155 Promoter architectures

156 C. difficile encodes a high number of which serve specialized functions, such as the

157 regulation of toxin production or sporulationσ factors. Consensus sequences have been proposed for 158 several of them including the vegetative SigA (Soutourina et al., 2013), the general stress response 159 SigB (Kint et al., 2019; Kint et al., 2017) , the major transition phase SigH (Saujet et al., 2011) and 160 the sporulation-specific SigK (Saujet et al., 2013) sigma factors, mostly based on comparative 161 transcriptome analyses of respective deletion strains and their wild-type strain. By applying 162 MEME-based searches upstream of all TSS we were able to refine published consensus sequences 163 . 2C) and substantially extended their network of regulated genes

164 for(Table all theseS3 and four S1). σ factors (Fig 165 Unsurprisingly, approximately half of the detected TSS (1,188/2,293) were associated 166 with a SigA-type promoter (Table S3), which include the previously determined TSS of sigH and 167 transcriptional regulator gene clnR (Saujet et al., 2011; Serrano et al., 2016; Woods et al., 2018). 168 The general stress factor SigB is another widespread sigma factor in gram-positive bacteria (van 169 Schaik and Abee, 2005). However, among Clostridia, C. difficile is the only species encoding an 170 annotated homologue. At the onset of stationary phase, SigB controls about 25% of all C. difficile 171 genes, primarily those involved in metabolism, sporulation and stress responses (Kint et al., 172 2017). We were able to extend the number of experimentally mapped SigB-associated TSS from 173 20 (Kint et al., 2017) to 277 (Table S3). Functional categories assigned to those genes are DNA 174 integration and recombination, transcriptional regulation, and cell wall turnover. Similarly, we 175 extend the SigH regulon to 69 genes, which previously included approx. 40 genes or operons,

6 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

176 respectively (Saujet et al., 2011). SigH is the key factor of transition phase and sporulation

177 initiation in C. difficile. Many genes associated withσ a SigH-type promoter in our analysis have 178 functions related to the biosynthesis of amino acids, secondary metabolites and antibiotics, in 179 addition to general metabolic pathways. Further, we identified 52 genes associated with a 180 promoter signature for sporulation-specific sigma factor SigK, including known genes of the SigK 181 regulon such as sleC and cdeC (Pishdadian et al., 2015). In addition, we found 39 genes with 182 alternative TSSs that were associated with two different promoter sequences. For example, RecA, 183 which controls the DNA damage response by homologous recombinational repair of damaged 184 DNA, is associated with a SigA- and SigB-type promoter supporting its role in stress responses 185 (Fig. 2D).

186 Global mapping of transcript ends

187 To map transcript 3’ termini, we adopted the RNAtag-Seq protocol (Shishkin et al., 2015), a 188 technique utilizing initial adapter ligation to exposed RNA 3’ ends. This approach yielded 2,042 189 experimentally determined transcript 3’ ends, which were assigned to one of the following classes 190 according to their genomic location: 3’ UTR (downstream of an annotated CDS or non-coding RNA 191 locus), 5’ UTR (between the TSS and the start codon of an annotated CDS), CDS (within a coding 192 sequence), orphan (downstream of an orphan TSS) and CRISPR (associated with CRISPR array) 193 (Fig. 3A). The majority of transcription termination events mapped to the region downstream of 194 annotated genes allowing us to generate a genome-wide map of 3’ UTRs (Table S4). Our data 195 reveal that 42% of all detected C. difficile 3’ UTRs downstream of an annotated CDS are >100 bp 196 in length (Fig. 3C), which resembles the 3’ UTR length distribution in Staphylococcus aureus (Ruiz 197 de los Mozos et al., 2013).

198 For the majority of 3’ UTR termination sites (75%) we identified a consensus motif that is 199 in agreement with known sequence features of Rho-independent terminators (Fig. 3C and Table 200 S4). This indicates a major role of Rho-independent termination in C. difficile and resembles 201 transcription termination in B. subtilis where the terminator protein Rho is not essential, 202 regulating transcription termination for only a few genes including rho itself (Ingham et al., 1999; 203 Nicolas et al., 2012; Quirk et al., 1993). Further, recent studies in Escherichia coli (E. coli) showed 204 that bidirectional transcription termination is a pervasive transcriptome feature in bacteria (Ju et 205 al., 2019). Accordingly, our analysis of transcription termination sites in C. difficile revealed many 206 overlapping termination events located between convergently transcribed genes, such as for 207 cwp19 and its convergent gene CDIF630_03029 or for the toxin gene toxA and the negative 208 regulator of toxin gene expression tcdC (Fig. 3D).

7 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

209 Small ORFs

210 High-resolution transcriptome maps allow the identification of ORFs that have been overlooked 211 in automated genome annotation. This includes so called small ORFs (sORFs) of usually 50 amino 212 acids or less, an emerging class of bacterial genes with an unfolding spectrum of new biological 213 functions (Garai and Blanc-Potard, 2020; Hemm et al., 2020). In many cases, they are predicted to 214 be membrane proteins, containing an alpha-helical transmembrane domain (Storz et al., 2014). 215 Focusing on oTSS in particular, we searched for novel ORFs based on the following criteria (i) 216 presence of a start and stop codon (ii) presence of a RBS within 15 bp upstream of the start codon 217 and (iii) sequence conservation in other Clostridioides strains. Based on this approach, we 218 identified 12 sORF candidates, seven of which are predicted to contain a transmembrane helix 219 (Table S2). Among the identified candidates six are toxins that are part of previously identified 220 C. difficile type-I toxin-antitoxins systems which were not annotated (Maikova et al., 2018; 221 Soutourina, 2019). Among the remaining sORFs four are high confidence candidates; one being a 222 conjugal transfer protein with annotation in other C. difficile strains; two candidates that each 223 have a Shine-Dalgarno (SD) -helical transmembrane domain; and one

224 sORF associated with a 160ntsequence long and5’ UTR a predicted region αharboring a c-di-GMP-I riboswitch. This 225 association with a cyclic di-GMP responsive riboswitch suggests a potential virulence-associated 226 function since c-di-GMP regulates not only motility and biofilm formation, but also toxin 227 production (Bordeleau et al., 2011; McKee et al., 2018; McKee et al., 2013; Peltier et al., 2015).

228 5’ UTRs and associated regulatory elements

229 Bacterial 5’ UTRs can influence gene expression in response to environmental signals, usually 230 through embedded cis-regulatory elements such as riboswitches, RNA thermometers and genetic 231 switches. Similar to other bacterial species, such as B. subtilis and Listeria monocytogenes, the 232 majority of 5’ UTRs in C. difficile range from 20 to 60 nucleotides in length (Fig. 4A, Table S1; 233 (Irnov et al., 2010; Sharma et al., 2010; Wurtzel et al., 2012). An aGGAGg motif that likely serves 234 as an RBS was detected in ~90% of these experimentally mapped 5’ UTRs (Fig. 4A, inlet). In 235 addition, we identified six leaderless mRNAs with a 5’ UTR of <10 nucleotides in length, including 236 the spoVAE gene within the tricistronic spoVACDE operon (Donnelly et al., 2016) indicating a 237 potential transcription of spoVAE that is uncoupled from the other two genes of the proposed 238 operon.

239 In addition to the average sized 5’ UTRs, a high number of genes are associated with 240 surprisingly long 5’ UTRs: 561 possess a 5’ UTR of >100 nucleotides, and 129 of those have a very 241 long 5’ UTR of >300 nucleotides. The latter include the 496 nt long 5’ UTR of flgB, which was 242 previously shown to contain both a c-di GMP-I riboswitch and a genetic switch that co-regulate 243 the expression of C. difficile flagella and toxin genes (Anjuwon-Foster and Tamayo, 2017). Further

8 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

244 analysis of these 5’ UTRs resulted in the identification of 77 Rfam-predicted riboswitch candidates 245 (Kalvari et al., 2018) (Fig. 4B). According to our RNAtag-Seq data, premature transcription 246 termination was evident in 57 of those 77 predicted riboswitches, indicating their ON state in 247 which transcription of the parental gene is repressed (Table S5). Interestingly, we detected a speF 248 riboswitch previously only documented in gram-negative alpha-proteobacteria associated with 249 CDIF630_01955, encoding a methyltransferase domain protein. As before, the associated 250 termination site (ON state) suggests that the riboswitch is functional, which is further supported 251 by a 220 nt northern blot signal in samples extracted from cultures grown in three different media 252 (Fig. 4D).

253 Additionally, we found two putative c-di GMP riboswitches to be associated with oTSSs: a 254 c-di GMP-I riboswitch in the vicinity of and opposite to a CRISPR12 array, associated with a 255 predicted sORF (see section New ORFs and Table S5) and a c-di GMP-II riboswitch immediately 256 downstream of prdC, encoding a subunit of the proline reductase enzyme. The latter is >900 nt 257 away from the closest coding region. No putative sORF was identified in this case, instead, the 258 associated regulated transcript might be an sRNA as published for three cobalamin riboswitches 259 in Enterococcus faecalis, Listeria monocytogenes and Streptococcus sanguinis each regulating an 260 sRNA (DebRoy et al., 2014; Mellin et al., 2014).

261 Besides riboswitch-associated premature termination, we observed 17 additional events 262 of putative premature transcription termination (PTT) in 5’ UTRs that lack similarity to conserved 263 riboregulators. A potential generation from mRNA processing is unlikely in these cases, since the 264 characteristic read enrichment in the untreated (TEX-) cDNA library, usually associated with 265 processing sites, is missing. One such PTT event is located in the 5’ UTR of the infC-rpmI-rplT 266 operon whose expression is controlled through an L20 leader via a transcription attenuation 267 mechanism in B. subtilis (Babina et al., 2018; Choonee et al., 2007). This ribosomal protein leader 268 autoregulatory structure is also found in other low-GC gram-positive bacteria (Rfam family 269 RF00558) but does not seem to be conserved on the primary sequence level in C. difficile. 270 Nevertheless, secondary structure prediction reveals extensive interactions between distant 271 bases that is reminiscent of the antiterminator conformation of the L20 leader from B. subtilis (Fig. 272 4C).

273 Further genes associated with such PTT events include several PTS systems (bglF, bglF1, 274 bglG3 and bglG4) that are known to be regulated by antiterminator proteins in B. subtilis (Fujita, 275 2009). Further, we also detected PTT events within the 5’ UTR of genes encoding transcriptional 276 regulators (CDIF630_00097, CDIF630_02384, CDIF630_02922), MDR-type ABC transporters 277 (CDIF630_03083, CDIF630_02847, CDIF630_03664) and aroF (Table S5). Northern blot 278 validation for a selection of candidates confirmed the RNA-seq predicted transcript sizes and

9 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

279 revealed larger bands in several cases that are likely to correspond to the full-length parental gene 280 or its degradation products (Fig. 4D).

281 RIP-seq identifies Hfq as a global RNA-binding protein in C. difficile

282 Our primary transcriptome analysis identified 42 novel transcripts that lack an internal open 283 reading frame, qualifying them as potential small regulatory RNAs (sRNA) (listed in Table S5). A 284 classification based on their genomic location (Fig. 5A) revealed the largest group to be 3’ UTR- 285 derived sRNAs (18), followed by those encoded cis-antisense either to a gene, another sRNA or 286 the 5’/3’ UTRs of coding sequences (13). In comparison, only few sRNAs were located in 287 intergenic regions (8) or derived from the 5’ UTR of mRNAs (3). Except for two transposon- 288 associated sRNAs (CDIF630nc_00004 and CDIF630nc_00069) and one located on a prophage 289 (CDIF630nc_00095), all of them are highly conserved among C. difficile strains, whereas sequence 290 conservation beyond C. difficile was extremely rare (Fig. S1, Table S7).

291 Northern blot validation of predicted sRNA candidates revealed that that most sRNA 292 candidates are expressed throughout exponential growth (Fig. 5B). About half of them are 293 downregulated after entry into stationary phase, while three candidates (CDIF630nc_00028 and 294 CDIF630nc_00089 and CDIF630nc_00105) show a marked accumulation. In agreement with the 295 increased abundance during stationary growth phase, CDIF630nc_00028 has a predicted 296 promoter sequences for SigB while CDIF630nc_00089 is associated with a SigH promoter 297 signature.

298 In many gram-negative model organisms such as Salmonella enterica and E. coli sRNAs 299 often require the RNA-binding protein Hfq to facilitate their interaction with target mRNAs (Hör 300 et al., 2020). However, whether Hfq functions as a central RNA-binding protein (RBP) in C. difficile 301 remains unknown. Therefore, we performed Hfq-immunoprecipitation followed by sequencing of 302 bound RNA species (RIP-seq) in a strain expressing C-terminally 3xFLAG tagged hfq (Hfq-FLAG) 303 along with its 5’ UTR under its native promoter to draft the spectrum of RNAs that are associating 304 with Hfq in vivo. Bacteria were grown in TY (tryptone yeast) medium to mid-exponential, late- 305 exponential and stationary phase (Fig. 6A) and subjected to the RIP-seq protocol (Fig. 6B). 306 Western blot validation using monoclonal FLAG antibody confirmed expression of Hfq throughout 307 all growth phases and specific enrichment of Hfq-FLAG from C. difficile 630 lysate (Fig. 6C). In 308 addition, co-purification of Hfq-bound sRNAs was confirmed by northern blot analysis of lysates 309 (L) and eluate (E) fractions from Hfq-FLAG and Hfq-ctrl bacterial cultures (Fig. 6C).

310 The majority of RNA species that co-immunoprecipitated with Hfq-FLAG were coding and 311 non-coding regions of mRNAs (CDS, 5’ UTRs and 3’ UTRs) (Fig. 6D). Interestingly, we observed 312 enrichment of many 5’ UTRs harboring riboswitches in addition to all but one type-I antitoxin 313 transcripts (Table S6). Importantly, the majority of putative sRNAs associated with Hfq (Fig. 6E, 10 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

314 see also Table S5). Among them was CDIF630nc_00070 (Fig. 6C), an sRNA previously shown to 315 bind to Hfq in vitro and in vivo (Soutourina et al., 2013). Interestingly, classifying Hfq-associated 316 sRNA with respect to their genomic location revealed that the majority of cis-antisense encoded 317 sRNAs in C. difficile were bound by Hfq (Fig. 6E). Profiling of Hfq-bound sRNAs revealed that the 318 majority was bound by Hfq across all three analysed growth phases with CDIF630nc_00070 and 319 CDIF630nc_00079 being the two top-enriched sRNAs (Fig. 6F, Fig S2A and Table S6). However, a 320 few sRNAs display a growth-phase dependent association with Hfq, such as CDIF630nc_00090 321 which is only enriched in stationary growth phase (Fig S2A).

322 Prediction of regulatory interactions for the novel Hfq-dependent sRNA AtcS

323 In addition to many sRNA candidates, we observed widespread binding of mRNA transcripts (CDS, 324 5’ UTRs and 3’ UTRs) (Fig. 6D, Table S6) who showed a much stronger growth-phase dependent 325 association with Hfq (Fig. 7A). In total, we identified 332 mRNAs as significantly enriched in our 326 Hfq pull- . Those mainly included genes involved in the biosynthesis of amino acids,

327 as well asdown membrane (log2 ≥ 2)transport, signal transduction and translation. Most interestingly, we also 328 found several mRNAs encoding for proteins important for virulence in C. difficile such as genes 329 involved in sporulation (spo0A, spoVB, oppB, sigG and sspA2), toxin production (tcdE), motility 330 (flgB, fliJ, flhB and cheW1), quorum sensing (luxS and agrB), the cell wall protein encoding gene 331 cwpV as well as several adhesins potentially involved in biofilm formation (CDIF630_03429, 332 CDIF630_01459 and CDIF630_03096) (Fig. 7B, Fig. S2B, Table S6).

333 To identify possible sRNA-mRNA interactions we applied the CopraRNA tool (Raden et al., 334 2018) to Hfq-associated sRNAs, focusing on interactions close to the RBS, since many sRNAs 335 regulate their target mRNAs through binding close to the RBS. By comparing the in silico predicted 336 potential target mRNAs with mRNAs that co-immunoprecipitated with Hfq in vivo, we readily 337 identified the 202 nt bona-fide intergenic sRNA CDIF630nc_00085 (Fig. 7C) as having an 338 interesting predicted target spectrum. For reasons explained below, we renamed the sRNA AtcS 339 (adhesin-targeting Clostridioides sRNA).

340 Northern blot validation of AtcS expression indicated sustained expression during 341 exponential growth that ceases in stationary phase, an effect that is more pronounced in glucose- 342 rich TYG medium as compared to TY and BHI (Fig. 7D). In agreement with the expression profiles, 343 AtcS is associated with a SigA promoter motif. In silico secondary structure prediction revealed a 344 moderately folded structure with an extended stem loop structure at the 5’ end, two smaller stem 345 loops in the center of the sRNA, and the Rho-independent terminator at the 3’ end.

346 CopraRNA predicts at least ten mRNAs as target candidates for AtcS that were also 347 enriched in the RIP-seq dataset (Fig. 7E&F). The corresponding predicted seed region in AtcS is 348 located in a single-stranded stretch adjacent to the Rho-independent terminator (Fig. 7C, red 11 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

349 circle) that base pairs with the region surrounding the SD sequence for most of the mRNA 350 candidates. Accordingly, aligning these mRNA target sites identified the following motif: AGGGG. 351 Targeting the region surrounding the SD sequence is a well described mechanism of sRNA- 352 mediated regulation. Perhaps the most common outcome of this interaction is inhibition of 353 translation by preventing ribosome binding to the mRNA (Hör et al., 2020).

354 Among the target candidates for AtcS, three are adhesins, of which two have homologs in 355 Staphylococcus epidermidis (CDIF630_03429 and CDIF630_01459) where they are involved in 356 surface adhesion and biofilm formation (Arrecubieta et al., 2007). Another predicted target is the 357 oligopeptide transporter oppB (CDIF630_00972), encoding a conserved oligopeptide permease 358 (Edwards et al., 2014) that was shown to influence sporulation initiation in C. difficile. Considering 359 the nature of the predicted AtcS interactions, this could provide a potential mechanistic link 360 explaining the increased biofilm formation and sporulation rate of a hfq mutant.

361 DISCUSSION

362 An online browser for easy access to the C. difficile transcriptome

363 In the present study, we have applied global approaches of bacterial RNA biology (Hor et al., 2018) 364 to capture the transcriptome architecture of C. difficile and unravel the scope of post- 365 transcriptional regulation in this important human pathogen. Through the combination of 366 unbiased experimental approaches to detect both transcriptional start sites and transcript 3’ ends, 367 we have bypassed the challenge of identifying novel non-coding regulators lacking conservation 368 to known RNA families. This is particularly important for Clostridia given their ancient origin and 369 the significant genomic differences even between C. difficile strains and other pathogenic 370 clostridia. The transcriptome maps are available for the scientific community through an open- 371 access online web browser. We intend to update our database whenever a refined genome 372 annotation becomes available, and to extend it by transcriptome profiles obtained under growth 373 in various infection-relevant conditions that will hopefully facilitate future genetic strategies.

374 Discovery of novel regulatory functions located in 5’ UTRs

375 Premature transcription termination (PTT) within 5’ untranslated regions were recently shown 376 to be widespread events in the Gram-positive bacteria B. subtilis, Listeria monocytogenes and 377 Enterococcus faecalis (Dar et al., 2016). Through our transcriptome-wide mapping of 3’ transcript 378 ends, we found 74 premature transcription termination events that were located in the 5’ UTRs 379 of mRNA transcripts. While 57 were associated with conserved riboswitches, the remaining 17 380 were lacking conservation to known families of ribo-regulators. Since such PTT events often have 381 regulatory functions, a response to unknown ligands or other unknown antitermination

12 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

382 mechanisms is likely. Interestingly, some of the detected PTT events reside in the 5’ UTRs of 383 multidrug-family type ABC transporters. Hence, characterizing their expression during exposure 384 of C. difficile to different antibiotic molecules will potentially reveal novel mechanisms underlying 385 antibiotic resistance in C. difficile. More broadly, mapping of 3 ends in C. difficile grown under

386 various growth conditions should reveal even more regulators ́ that are not active during the 387 standard laboratory growth conditions applied in this study. Such an approach recently 388 discovered antibiotic-responsive transcription termination events upstream of ABC transporters 389 and transcriptional regulators in B. subtilis, Listeria monocytogenes and Enterococcus faecalis (Dar 390 et al., 2016) and demonstrates the widespread presence of such mechanisms as well as their 391 potential importance for bacterial physiology.

392 A large Hfq network in a Gram-positive bacterium

393 Our knowledge of sRNA-based gene regulation in gram-positive bacteria, particularly in C. difficile, 394 is lagging behind that of gram-negative model organisms e.g. Salmonella enterica and E. coli. 395 However, it is undisputed that sRNAs are central players in the regulation of physiological and 396 virulence pathways in gram-positive bacteria, such as Listeria monocytogenes, B. subtilis and 397 Staphylococcus species. For example, the sRNA RsaE, that is conserved in the order Bacillales, has 398 been shown to coordinate central carbon and amino acid metabolism in Staphylococcus aureus 399 (Bohn et al., 2010; Rochat et al., 2018) and also to promote biofilm formation in Staphylococcus 400 epidermidis by supporting extracellular DNA release and the production of polysaccharide- 401 intercellular adhesin (PIA) (Schoenfelder et al., 2019). However, the function of Hfq in facilitating 402 the interactions between sRNA regulators and their target mRNAs in Gram-positive bacteria 403 remains elusive, since most of the known sRNAs seem to exert their regulatory activity 404 independent of this global RNA-binding protein (Pitman and Cho, 2015).

405 Previous efforts using in silico applications (Chen et al., 2011) and RNA-sequencing 406 (Soutourina et al., 2013) have predicted >100 sRNA candidates in C. difficile, implying the 407 existence of a large post-transcriptional network. Here, we provide experimental evidence for the 408 existence of 42 sRNAs (that we consider high confidence candidates) which either originate from 409 independent transcriptional start sites or are processed from untranslated regions of mRNAs. 410 This number is at the lower spectrum of estimated sRNAs for various bacterial species (Adams 411 and Storz, 2020). However, considering that our analysis was done in only two different growth 412 phases of C. difficile cultured in rich media, this number is likely to increase in the future. 413 Reflecting this assumption, most of our sRNAs were expressed throughout exponential growth 414 but decreased in stationary phase. Therefore, mapping the transcriptome under defined stress 415 conditions will likely extend the sRNA repertoire of C. difficile by novel candidates that are, for

13 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

416 example, under the control of stress-induced transcriptional regulators and therefore not 417 expressed under the conditions that were assayed in this work.

418 Small regulatory RNAs are classified according to the genomic location they originate 419 from. To date, the largest number of experimentally characterized sRNAs are those located in 420 intergenic regions that are transcribed from their own promoter and typically have a Rho- 421 independent terminator. However, especially sRNAs derived from the 3’ UTR of mRNAs constitute 422 an expanding class that is gaining increasing attention as more examples are being functionally 423 characterized (Adams and Storz, 2020; De Mets et al., 2019; Hoyos et al., 2020; Miyakoshi et al., 424 2019). Interestingly, only a small portion of our sRNA candidates originates from intergenic 425 regions (8 of 42) whereas the largest class were, in fact, sRNAs derived from 3’ UTRs (18 of 42).

426 Many of the 3’ UTR-derived sRNAs identified in our study were associating with Hfq in 427 vivo suggesting they could act in trans to potentially cross-regulate other mRNA targets. One 428 interesting example is CDIF630nc_00006 which is generated by processing from the 3’ UTR of 429 cbiO, encoding the ATP-binding protein of a cobalt-specific ABC-transporter. Cobalt is mainly 430 incorporated into vitamin B12 which in turn is an important enzyme cofactor in a variety of 431 biological processes in C. difficile (Shelton et al., 2020). A prominent one is the utilization of 432 ethanolamine, an important carbon source for C. difficile during colonization of the gut whose 433 availability impacts disease outcome (Nelson et al., 1980).

434 Another abundant class of sRNA (13 of 42) were cis-encoded antisense RNAs. 435 Interestingly, this does not reflect genome-wide antisense transcription which only accounted for 436 approx. 6% of TSS events in C. difficile. Comparative transcriptomics have revealed that bacterial 437 antisense transcription is mainly the product of transcriptional noise and directly correlates with 438 genomic AT content (Lloréns-Rico et al., 2016). With a GC content of only 27% in C. difficile this 439 result raises the question whether our annotated antisense RNAs are functional regulators. 440 However, the majority of cis-antisense encoded sRNAs (10 of 13) were bound by Hfq which 441 suggests a regulatory function for these transcripts. If expressed in sufficient amounts they can 442 regulate the expression of the gene encoded on the opposite strand, an interaction that is not 443 considered to require the presence of Hfq in other studied bacteria due to extensive sequence 444 complementarity (Irnov et al., 2010). Accordingly, we found none of these mRNAs associated with 445 Hfq in vivo. This could indicate that there are additional trans-encoded target mRNAs regulated 446 through these antisense sRNAs for which binding would need to be facilitated by Hfq as a 447 consequence of imperfect complementarity.

448 Implications of Hfq associations for C. difficile virulence

449 A recent characterization of a Hfq deletion strain revealed a hyper-sporulation phenotype of the 450 mutant in comparison to the wild type strain suggesting that sRNAs participate in the repression 14 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

451 of sporulation (Maikova et al., 2019). Among the Hfq-bound mRNAs we found the sporulation- 452 specific transcriptional regulator, sigG, and the master regulator of sporulation, spo0A, associated 453 with Hfq in vivo. In addition, the oligopeptide transporter oppB, which regulates the initiation of 454 sporulation, was enriched in the Hfq pulldown. The latter was also among the COPRA-predicted 455 target mRNAs of the bona fide intergenic sRNA AtcS. Together, these mRNAs provide a first 456 mechanistic explanation for the observed hyper-sporulation phenotype of a hfq mutant (Maikova 457 et al., 2019). Until recently, only protein regulators were considered to be important for the 458 sporulation process. However, several studies in B. subtilis have identified sRNA candidates to be 459 controlled by sporulation-specific sigma factors including SigG, SigK, SigF and SigE, suggesting an 460 involvement of RNA players in the regulation of sporulation (Marchais et al., 2011; Schmalisch et 461 al., 2010; Silvaggi et al., 2006). Likely due to the standard growth conditions used for our 462 experimental transcriptome annotation, we did not find any sRNA candidates that are transcribed 463 from a sporulation-specific promoter.

464 Moreover, we found several mRNAs encoding for proteins involved in motility (flgB, fliJ, 465 flhB and cheW1), the cell wall protein encoding gene cwpV as well as several adhesins potentially 466 involved in biofilm formation (CDIF630_03429, CDIF630_01459 and CDIF630_03096) that could 467 explain the observed phenotypes of a Hfq knock-down, including reduced motility, increased 468 sensitivity to stresses and biofilm formation (Boudry et al., 2014). Some of the identified Hfq- 469 bound mRNAs, such as spo0A and cwpV, were also found to be differentially regulated in the same 470 Hfq knock-down strain, further confirming them as direct potential targets of sRNA regulation in 471 C. difficile.

472 In total, we identified 330 mRNAs associating with Hfq in vivo. This would imply that about 473 10% of the genome is under direct post-transcriptional regulation. However, our RIP-seq based 474 identification of Hfq-associated mRNAs includes the sigma factors sigA2 and the sporulation sigma 475 factor sigG as well as >20 transcriptional regulators and several two component systems. 476 Therefore, it is reasonable to assume that sRNA-based regulation could impact gene expression 477 to a much greater extent than suggested by the 330 mRNAs enriched in our analysis.

478 Identification of conserved sRNA with members outside the class Clostridia

479 Most of the sRNAs identified in this study are exclusively conserved within the species C. difficile. 480 This lack of broader sequence conservation is probably not surprising in light of the recent re- 481 classification of C. difficile that placed the species in the new taxonomic family 482 Peptostreptococcaceae (Lawson et al., 2016). However, the evolution of sRNAs is far less 483 understood than the evolution of proteins which can be conserved over long phylogenetic 484 distances whereas individual sRNAs tend to display narrow conservation ranges restricted to the 485 genus or family level (Jose et al., 2019). Interestingly, in some cases, sRNAs from distant phyla

15 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

486 share the same function and target mRNAs despite a lack of primary sequence conservation. One 487 such example is the Fur-regulated sRNA FsrA in B. subtilis, that regulates mRNA targets involved 488 in iron transport and metabolism (Gaballa et al., 2008; Pi and Helmann, 2017) in a very similar 489 way to RyhB in Enterobacteriaceae (Masse and Gottesman, 2002; Masse et al., 2005). The genome 490 of C. difficile 630 lacks any sequence conservation to FsrA but the transcriptomic response of a 491 C. difficile 630 Fur mutant involves the down-regulation of similar genes and pathways as in 492 B. subtilis (Berges et al., 2018). Therefore, future analyses of transcriptional responses to iron 493 starvation and other defined stress conditions might reveal novel sRNAs in C. difficile that have 494 functionally related members in other bacteria.

495 Although most sRNAs were only present in C. difficile, two sRNAs (CDIF630nc_00069 & 496 CDIF630nc_00001), showed a remarkably high sequence conservation outside the class 497 Clostridia, indicating that they serve in cellular pathways that are conserved across multiple 498 species (Frohlich et al., 2016). In Gram-positive organisms such core sRNA are not well 499 characterized but there are a few examples, such as SR1, a dual function sRNA with broad 500 conservation within Bacillales (Gimpel et al., 2012).

501 The first sRNA, CDIF630nc_00069, is associated with a putative conjugative transposon 502 element. As a result, this sRNA is absent in some C. difficile strains. Hence, while CDIF630nc_00069 503 is present in C. difficile 630 (RT012) and the epidemic and hypervirulent strain R20291 (RT027), 504 the sRNA is absent in the non-epidemic strain CD196 of the same RT027 ribotype (Stabler et al., 505 2009). Instead, CDIF630nc_00069 is sporadically found in distant bacterial lineages including 506 Streptococcus pneumoniae and Streptococcus pyogenes (Fig. 5C and Table S10). Taken together, 507 CDIF630nc_00069 seems to be part of a mobile genomic element that has spread amongst distant 508 members of the phylum Firmicutes.

509 The second sRNA, CDIF630nc_00001, is the first experimentally validated member of the 510 RaiA family of structured sRNA that was discovered by bioinformatic approaches (Weinberg et 511 al., 2017), therefore, we rename it RaiA. RaiA RNAs are found in Firmicutes and Actinobacteria 512 and share conserved nucleotides as well as secondary structures. In both phyla, they often occur 513 upstream of the raiA gene that encodes for a stress‐inducible ribosome‐inactivating protein 514 (Agafonov and Spirin, 2004). In C. difficile strains the RaiA RNA resides in the 3’ UTR of 515 CDIF630_00250, encoding a putative phosphoribosyl transferase, whereas in other Clostridium 516 species including C. botulinum, C. perfringens, C. tetani and C. sporogenes the sRNA is located in the 517 commonly observed raiA 5’ UTR region. Our RNA-seq and RNAtag-Seq data revealed transcription 518 of RaiA starting from a dedicated TSS and two terminators resulting in two transcripts of different 519 lengths both of which could be detected by Northern Blot analysis (Fig. 5B).

520 Interestingly, neither CDIF630nc_00069 nor CDIF630nc_00001 was associating with Hfq 521 in vivo. This might suggest the potential existence of alternative RNA-binding proteins in 16 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

522 C. difficile, similar to Salmonella or E. coli where a second global RBP, ProQ, was recently 523 characterized (Holmqvist and Vogel, 2018; Melamed et al., 2020; Smirnov et al., 2016). Potential 524 candidates in Gram-positive bacteria exist, such as the KhpA/B heterodimer identified in 525 Streptococcus pneumoniae which was shown to associate with several RNA classes leading to its 526 classification as a global RNA chaperone (Zheng et al., 2017). Further, a recent Grad-seq analysis 527 of cellular RNA-protein complexes in Streptococcus pneumoniae detected a KhpA/B complex that 528 co-sedimented with tRNAs and intergenic sRNA transcripts (Hor et al., 2020). Given that we are 529 only beginning to understand the true number and functions of sRNA and protein players involved 530 in RNA-based gene regulation in C. difficile, there is great potential to discover previously 531 unknown RNA-binding proteins in this important human pathogen.

532 METHODS

533 Bacterial strains and growth conditions

534 A complete list of C. difficile and E. coli strains that were used in this study is provided in Table S8.

535 C. difficile cultures were routinely grown anaerobically inside a Coy chamber (85% N2, 10% H2

536 and 5% CO2) in Brain Heart Infusion (BHI) broth or on BHI agar plates (1.5% agar) unless 537 indicated otherwise. When necessary, antibiotics were added to the medium at the following 538 concentrations: thiamphenicol 15 , cycloserine 250 . E. coli cultures were propagated

539 aerobically in Luria-Bertani (LB) brothμg/ml (10 g/l tryptone, 5μg/ml g/l yeast extract, 10 g/l NaCl) or on LB 540 agar plates (1.5% agar) supplemented with chloramphenicol (20 as appropriate. E. coli

541 strain Top10 (Invitrogen) was used as a recipient for all cloning procedures,μg/ml) and E. coli CA434 542

543 (HB101of plasmids carrying into C. the difficile IncPβ 630. conjugative plasmid R702) was used as donor strain for conjugation

544 Plasmid construction

545 All plasmids and DNA oligonucleotides that were used are listed in Tables S9 and S10, 546 respectively. hfq including its native 5’ UTR and native promoter was amplified from C. difficile 547 630 using either FFO-122 and FFO-136 or FFO-122 and FFO-207. The resulting fragments were 548 cloned into SacI/BamHI-digested pRPF185 (Fagan and Fairweather, 2011), generating pFF-10, 549 encoding hfq including 5’ UTR and native promoter, and pFF-12, encoding a C-terminally 3X- 550 FLAG-tagged hfq including 5’ UTR and native promoter, respectively.

551 Strain construction

552 pFF-10 and pFF-12 were transformed into chemically competent E. coli TOP10 according to 553 standard procedures (J. Sambrook, 1989). Both strains (FFS-34, harboring pFF-10 and FFS-46,

17 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

554 harboring pFF-12) were used for cloning and plasmid propagation. For conjugation purposes, 555 both plasmids were transformed in E. coli

556 R702), resulting in FFS-36 and FFS-48 respectively.CA434 (HB101 Conjugation carrying was the performed IncPβ conjugative according plasmid to Kirk 557 and Fagan, 2016 (Kirk and Fagan, 2016). In short: 200 C. difficile 630 overnight cultures were

558 incubated at 37 °C for 2 min. Simultaneously, 1 ml ofμl overnight of E. coli conjugant donor culture 559 (FFS-36 and FFS-48) was harvested by centrifugation at 4000 x g for 2 min. E. coli pellets were 560 then transferred into the anaerobic workstation and gently resuspended in pre-incubated 200

561 C. difficile 630 culture. Following resuspension, the cell suspension was pipetted onto well-dried,μl 562 non-selective BHI agar plates (10 × 10 , allowed to dry and incubated for 8 h at 37 °C.

563 Growth was harvested using 900 μl spots) 564 either cycloserine (control), or cycloserineμl of TY broth, and serially thiamphenicol, diluted and to spreadselect foron platestransconjugants. containing 565 Plates were incubated for between 24 and 72 h, until colonies were apparent. Conjugation 566 resulted in strain FFS-38, harbouring pFF-10 and strain FFS-50, harbouring pFF-12, which were 567 used for RIP-seq analysis.

568 Total RNA extraction

569 Total RNA was extracted using the hot phenol protocol. Bacterial cultures were grown to the

570 desired OD600, mixed with 0.2 volumes of STOP solution (95% ethanol, 5% phenol) and snap- 571 frozen at -80 °C if not directly processed. The bacterial solution was centrifuged for 20 min, 572 4500 rpm at 4 °C and the supernatant was completely discarded. Cells were suspended in 600

573 of 10 mg/ml lysozyme in TE buffer (pH 8.0) and incubated at 37 °C for 10 min. Next, 60 μl 574 w/v SDS was added and everything mixed by inversion. Samples were incubated in a waterμl of 10%bath 575 at 64 °C, 1-2 min before adding 66 M NaOAc, pH 5.2. Next, 750 -Aqua

576 phenol) was added, followed by incubationμl 3 for 5 min. at 64 °C. Samplesμl were acid briefly phenol placed (Roti on ice 577 to cool before centrifugation for 15 min, 13,000 rpm at 4 °C. The aqueous layer was transferred 578 into a 2 ml phase lock gel tube (Eppendorf), 750

579 everything centrifuged for 12 min, 13,000 rpm atμl room chloroform temperature. (Roth, For#Y015.2) ethanol was precipitation, added, and 580 the aqueous layer was transferred to a fresh tube, 2 volumes of 30:1 mix (EtOH:3 M NaOAc, 581 pH 6.5) was added and incubated overnight at -20 °C. Precipitated RNA was harvested by 582 centrifugation, washed with cold 75% v/v ethanol and air-dried. DNA contaminations were 583 removed by DNase treatment and the RNA was re-extracted using a single phenol-chloroform 584 extraction step. Purified RNA was resuspended in 50 µl RNase-free water and stored at -80 °C.

585 Library preparation for differential RNA-seq (dRNA-seq)

586 Library preparation for dRNA-seq was accomplished by Vertis Biotechnology AG. In brief, total 587 RNA was analyzed on a Shimadzu MultiNA microchip electrophoresis system. 23S/16S ratio for

18 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

588 all samples was 1.3. RNA was fragmented via ultrasound (4 pulses a 30 sec, 4 °C) and subsequently 589 treated with T4 Polynucleotide Kinase (NEB). Half of the samples were then treated with 590 terminator exonuclease (TEX) for dRNA-seq.

591 For the cDNA synthesis, the RNA fragments were poly(A)-tailed and 5’PPP structures were 592 removed with RNA 5’ Polyphosphatase (Epicentre). The RNA sequencing adapter with the 593 barcodes were ligated to the 5’-monophosphate of the fragments. First strand cDNA was 594 synthesized with an oligo(dT)-adapter primer and M-MLV reverse transcriptase. Amplification of 595 cDNA was done via PCR to an approximate amount of 10-20 ng/µl. cDNA was purified with the 596 Agencourt AMPure XP kit (Beckman Coulter Genomics) and analyzed with Shimadzu MultiNA 597 microchip electrophoresis. Equimolar amounts of the samples were pooled for sequencing. cDNAs 598 had a size between 200 and 550 bp. The library pool was fractionated via differential clean-up 599 with the Agencourt AMPure kit. The cDNA pool was checked with capillary electrophoresis as 600 stated above. Libraries were sequenced on an Illumina NextSeq 500 system using 75 bp read 601 length.

602 Library preparation for RNAtag-Seq protocol

603 Total RNA quality was checked using a 2100 Bioanalyzer with the RNA 6000 Nano kit (Agilent 604 Technologies) and rRNA was detected. The RIN for all samples was >6.5. Equal amounts of 605 samples (~500 ng) were used for the preparation of cDNA libraries with the RNAtag-Seq protocol 606 as previously published by Shishkin et al. (Shishkin et al., 2015) with minor modifications.

607 Briefly, the RNA samples were fragmented with FastAP buffer at 94 °C for 3 min and were 608 dephoshorylated using the FastAP enzyme (Thermo Scientific) at 37 °C for 30 min followed by 609 bead purification with 2 volumes of Agencourt RNAClean XP beads. Fragmentation profiles were 610 checked using RNA 6000 pico kit (Agilent) with a 2100 Bioanalyzer. The RNA fragments were 611 ligated with 3´-barcoded adaptors at 22 °C for 1 h and 30 min using T4 RNA Ligase (NEB). 612 Barcoded RNA samples were pooled together, the ligase was inhibited by RTL buffer (Qiagen 613 RNeasy Min Elute Cleanup Kit) and purified with RNA Clean & Concentrator-5 column (Zymo). 614 rRNA were depleted from pools using Ribo-Zero (Bacteria) Kit (Illumina) and purified with RNA 615 Clean & Concentrator-5 column (Zymo). The profiles before and after rRNA depletion were 616 analysed with RNA 6000 pico kit (Agilent) with a 2100 Bioanalyzer. The rRNA depleted RNA were 617 reverse transcripted and the first strand of the cDNA were synthsized using a custom AR2 oligo 618 (Sigma-Aldrich) and the AffinityScript multiple temperature cDNA synthesis kit (Agilent) at 55 °C 619 for 55 min. The RNA was degraded with 1 M NaOH at 70 °C for 12 min, the reactions were 620 neutralized with Acetic acid and were purified with 2 volumes of MagSi-NGSPREP Plus beads 621 (AMSBIO). A second 3Tr3 adaptor were ligated to the cDNA using T4 RNA Ligase (NEB) overnight 622 at 22 °C followed with two bead purification steps with 2 volumes of MagSi-NGSPREP Plus beads

19 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

623 (AMSBIO). A PCR enrichment test was performed in order to determine the number of PCR cycles 624 that are nesessery for each pool followed by a bead purification step and a QC with DNA HS kit 625 (Agilent). Then, the final PCR was performed with the number of cycles determined from the 626 previous step using P5 and P7 primers. Two sequential bead purification steps were performed 627 with 1.5 and 0.7 respectively.

628 Libraries were quantified with the Qubit 3.0 Fluometer (ThermoFisher) and the library 629 quality and size distribution (~480 bp peak size) was checked using a 2100 Bioanalyzer with the 630 High Sensitivity DNA kit (Agilent). Sequencing of pooled libraries, spiked with 5% PhiX control 631 library, was performed with ~8 million reads / sample in single-end mode on the NextSeq 500 632 platform (Illumina) with the High Output Kit v2.5 (75 Cycles).

633 We noticed a strong enrichment of reads mapping to native 3’ ends of transcripts in 634 C. difficile 630 libraries that were prepared for the RNAtag-Seq protocol. We cannot fully explain 635 this observation but we suspect a combination of inefficient RNA fragmentation and native 3’ end 636 stability of transcripts to contribute to this outcome for RNAtag-Seq libraries prepared from 637 C. difficile 630 total RNA.

638 Northern blotting

639 RNA samples were separated on a denaturing 6% polyacrylamide gel in TBE buffer containing 640 7 M urea. Gels were transferred onto Hybond+ membranes (GE Healthcare Life Sciences) at 4 °C 641 with 50 V (~100 W) for 1 h. For hybridization with P32-labeled DNA oligonucleotides, membranes 642 were incubated over night at 42 °C in Roti® Hybri-Quick Buffer (Roth). Membranes were washed 643 three times with decreasing concentrations of SSC buffer (5 x, 1 x and 0.5 x) before imaging on a 644 Typhoon FLA 7000 phosphor imager.

645 Western blotting

646 To verify the expression and successful pulldown of FLAG-tagged Hfq, 9 µl lysate and 50 µl of 647 resuspended beads were mixed with 81 µl and 50 µl 1X protein loading dye respectively and 648 boiled for for 5 min at 98 °C. Following incubation, 20 µl of each sample was loaded and separated 649 on a 15% SDS–polyacrylamide gel followed by transfer of proteins to a PVDF membrane. For 650 detection of FLAG-tagged proteins, the membrane was blocked in TBS-T with 5% milk powder for 651 1 h at room temperature and washed 3 x in TBS-T for 10 min. Subsequently the membrane was 652 incubated over night at 4 °C with anti-FLAG antibody (Sigma) diluted 1:1,000 in TBS‐T with 3% 653 BSA and washed again 3 x in TBS-T for 10 min. Following the last washing step the membrane was 654 incubated for 1 h at room temperature with anti‐mouse‐HRP antibody (ThermoScientific) diluted 655 1:10,000 in TBS‐T with 3% BSA and finally washed 3 x in TBS‐T for 10 min before adding ECL 656 substrate for detection of HRP activity using a CCD camera (ImageQuant, GE Healthcare).

20 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

657 Hfq co-immunoprecipitation and RNA-sequencing (RIP-seq)

658 Overnight cultures of FFS-38 and FFS-50 were prepared in biological duplicates in buffered TY

659 and used for inoculation of pre-cultures. Once an OD600 of 0.5 was reached, each replicate was 660 used for inoculation of three separate flasks. Main cultures were grown until either mid-

661 exponential (OD600=0.5), late-exponential (OD600=0.9) or stationary (3 h post entry) growth 662 phase. For each condition, 50 OD were harvested by centrifugation for 20 min at 4000 rpm at 4 °C,

663 snap frozen and resuspended in 800 µl lysis buffer (20 mM TRIS pH 8.0, 1 mM MgCl2, 150 mM KCl, 664 1 mM DTT). Subsequently, each sample was mixed with 1 µl DNase I (Fermentas, 1 U/µl) and 665 800 µl of 0.1 mm glass beads and lysed in a RETSCH's Mixer Mill (30 Hz, 10 min, 4°C). Bacterial 666 lysates were cleared by centrifugation for 10 min at 14,000 x g and 4 °C. Approximately 900 µl 667 supernatant was transferred to a new tube and incubated with 25 µl (1/2 * OD in µl) of mouse- 668 anti-FLAG antibody (clone M2, Merck/Sigma- °C for 45 min (rocking).

669 Immunoprecipitation of FLAG-tagged Hfq was performedAldrich #F1804) by incubating at 4 each sample with 75 µl 670 pre-washed (3 x resuspended in 1 ml lysis buffer and centrifuged at 10,000 rpm for 1 min) protein 671 A sepharose beads (Merck/Sigma- min at 4 °C (rocking). Beads and

672 captured proteins were washed 5 x Aldrichwith 500 #P6649) µl lysis buffer for 45 (mixed by inversion and centrifuged 673 at 10,000 rpm for 1 min) and resuspended in 500 µl lysis buffer. Finally, RNA, co- 674 immunoprecipitated with antibodies and protein A sepharose beads, was eluted by adding the 675 same volume of Phenol:Chloroform:Isoamylalcohol (25:24:1, pH 4.5, Roth). The solution was 676 mixed for 20 s, transferred to PLG tubes (Eppendorf) and incubated at room temperature for 677 3 min. Following centrifugation (30 min, 15,200 x g, 15 °C) the aqueous phase was transferred to 678 new tubes and RNA was precipitated over night at -20 °C by adding 30 µg Glycoblue and 800 µl 679 isopropanol. Precipitated RNA was pelleted (centrifugation for 45 min, 15,200 rpm, 4 °C), washed 680 with 80% and 100% ethanol (centrifugation for 10 min, 15,200 rpm, 4 °C) and resuspended in 681 15,5 µl nuclease-free water (65 °C, 1 min, 600 rpm). For DNase treatment, purified RNA was 682 incubated with 2 µl DNase I, 0.5 µl RNase inhibitor and 2 µl 10 x DNase buffer for 30 min at 37 °C. 683 To each sample, 100 µl nuclease free water was added and DNase treated RNA was purified by 684 Phenol:Chloroform:Isoamylalcohol treatment, as described before. Precipitation was performed 685 overnight at -20 °C by mixing the samples with 3 x the volume of sodium-acetate/ethanol (1:30), 686 followed by centrifugation and washing with 80% and 100% ethanol. Finally, air-dried pellets 687 were resuspended in 20 µl nuclease-free water (65 °C, 1 min, 600 rpm).

688 RNA quality was controlled using a 2100 Bioanalyzer and the RNA 6000 Pico kit (Agilent 689 Technologies). Since rRNA was detectable, the RIN for all samples was >7.5. Equal amounts of RNA 690 (~50 ng) were used for preparation of cDNA libraries with the NEBNext Multiplex Small RNA 691 Library Prep kit for Illumina (NEB) with some minor changes to the manufacturers’ instructions. 692 RNA samples were fragmented with Mg2+ at 94 °C for 2 min, 45 s using the NEBNext Magnesium

21 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

693 RNA Fragmentation Module (NEB) followed by RNA purification with the Zymo RNA Clean & 694 Concentrator kit. Fragmented RNA was dephosphorylated at the 3’ end, phosphorylated at the 695 5’ end and decapped using 10 U T4-PNK +/-, 40 nmol ATP and 5 U RppH respectively (NEB). After 696 each enzymatic treatment, RNA was purified with the Zymo RNA Clean & Concentrator kit. The 697 RNA fragments were ligated for cDNA synthesis to 3’ SR adapters and 5’ SR adapters and diluted 698 1:5 with nuclease-free water before use. PCR amplification was performed to add Illumina 699 adaptors and indices to the cDNA for 16 cycles, using 1:5 diluted primers. Barcoded DNA Libraries 700 were purified using magnetic MagSi-NGSPREP Plus beads (AMSBIO) at a 1.8 ratio of beads to 701 sample volume. Finally, libraries were quantified with the Qubit 3.0 Fluorometer (ThermoFisher) 702 and the library quality and size distribution (~230 bp peak size) was checked using a 2100 703 Bioanalyzer with the High Sensitivity DNA kit (Agilent). Sequencing of pooled libraries, spiked 704 with 5% PhiX control library, was performed with ~5 million reads / sample in single-end mode 705 on the NextSeq 500 platform (Illumina), using the High Output Kit v2.5 (75 cycles). Demultiplexed 706 FASTQ files were generated with bcl2fastq2 v2.20.0.422 (Illumina). Reads were trimmed for the 707 NEBNext adapter sequence using Cutadapt version 2.5 with default parameters. In addition, 708 Cutadapt was given the –nextseq-trim=20 switch to handle two colour sequencing chemistry. 709 Reads that were trimmed to length=0 were discarded.

710 The READemption pipeline version 4.5 (Förstner et al., 2014) with segemehl version 0.2 711 (Otto et al., 2014) was used for read mapping and calculation of read counts. Reads were mapped 712 on CP010905.2 with additional annotations for ncRNAs, 3’ UTRs and 5’ UTRs. Normalization and 713 enrichment analysis was done with the edgeR version 3.28.1 (Robinson et al., 2010). To calculate 714 normalization factors the trimmed means of M values (TMM) method was used between flag- 715 tagged and non-flag-tagged Hfq libraries of each growth phase individually. All features that 716 considered more than 10 read counts in at least two replicates were included in the analysis. Only 717 features with a Benjamini-Hochberg corrected P-value of ≤0.1 and a log2 fold-

718 considered as significantly enriched. change of ≥2 were

719 Annotation of TSS, TTS and sRNAs

720 For the prediction of transcriptional start sites (TSSs) and small regulatory RNAs (sRNAs), we 721 used the modular, command-line tool ANNOgesic (Yu et al., 2018). All parameters were kept at 722 the default setting, unless stated otherwise. The TSS prediction algorithm was trained using 723 parameters that were derived from manual curation of the first 200 kb of the genome. Secondary 724 TSSs were excluded if a primary TSS was within seven nucleotides distance. For annotation of 5’ 725 untranslated regions (5’ UTRs) the default settings for maximum length was corrected to 500 nt.

726 For the prediction of sRNAs, we used the default settings, that is, the transcript must have 727 a predicted TSS or processing site (PS) at their 5’ end, form a stable secondary structure

22 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

728 (calculated with RNAfold from Vienna RNA package), and a length between 30 to 500 nt. For 729 5’ UTR-derived sRNAs the transcripts must have a PS at their 3’ end. For 3’ UTR-derived sRNA the 730 transcript can either have a TSS or PS at their 5’ end and share its terminator with the parental 731 gene.

732 Prediction of transcriptional termination sites (TTSs) was accomplished manually. 733 Therefore, read coverage tracks of RNAtag-Seq libraries were visualized with Integrative 734 Genomics Viewer (Robinson et al., 2011) and scanned for enriched regions. Termination sites 735 were defined as the first position with less than half of the maximal reads of the entire enriched 736 region.

737 Prediction of promoter and terminator motifs, novel ORFs

738 Motif search and generation of sequence logos was accomplished with MEME version 5.1.1 (Bailey 739 and Elkan, 1994) and CentriMO version 5.1.1 (Bailey and Machanick, 2012). For determination of 740 a sigA promoter motif, 100 nt upstream of each TSS was extracted and uploaded to MEME. Motifs 741 to be found by MEME was set to 10 and only the given strand was searched. All other settings 742 were left at default. The resulting sigA motif matrix was then uploaded to CentriMo along with the 743 extracted promoter regions. Again, only the given strand was searched and all other settings were 744 left at default. Based on the CentriMo output, those promoter sequences harboring a sigA 745 promoter motif were uploaded to MEME ones more to refine the sigA promoter motif. In contrast 746 to sigA, previously published promoter motifs harboring either a sigB, sigH or sigK promoter 747 motifs were uploaded for the initial MEME search instead of extracted promoter sequences 748 generated in this study. The input sequences were based on the following studies: sigB (Kint et al., 749 2017); sigH (Saujet et al., 2011); sigK (Saujet et al., 2013). The thereby generated motif matrices 750 were uploaded to CentriMo together with our extracted promoter sequences. Those sequences 751 harboring a sigB, sigH or sigK motif were uploaded to MEME to generate a refined promoter motif 752 based on our input sequences. Following the MEME search, some sequences were excluded 753 manually based on the following criteria: promoter location within the uploaded sequence has to 754 be at the 3’ end; at least one G within 3 nucleotides next to the -10 box (only sigB); G at position 755 25, 26 or 27 (only sigH); ATA at position 23-25 and at least one A and C between position 2-7 (only 756 sigK). Finally, the manually curated promoter sequences were uploaded to MEME to generate the 757 final promoter motif.

758 For 3’ UTR motif search 15 nt downstream of all termination sites were extracted to 759 search for the presence of the common poly-U stretch. Command line version of MEME was 760 executed with default options except for the minimal motif length, which was set to 4. Only the 761 given strand was used for the search.

23 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

762 Prediction of novel ORFs was accomplished manually. In order to identify novel ORFs we 763 focused on oTSS, - TSS not associated with any coding or non-coding sequence. To avoid any miss- 764 annotation we applied the following criteria: (i) presence of a start and stop codon (ii) presence 765 of a ribosome binding site within 15 bp upstream of a start codon (iii) sequence conservation in 766 other Clostridioides strains.

767 Prediction of RNA folding and sRNA target mRNAs

768 Secondary structures of ncRNAs were predicted with the RNAfold WebServer (Lorenz et al., 2011) 769 and visualized with VARNA (Darty et al., 2009). To predict folding of TTS the sequence 60 nt 770 upstream and 10 nt downstream of the termination site was extracted. RNAfold version 2.2.7 of 771 the Vienna RNA Package 2 was used with default settings to predict secondary structures.

772 Potential sRNA-target interactions were predicted using CopraRNA (Raden et al., 2018; 773 Wright et al., 2014; Wright et al., 2013). Target prediction was performed for all sRNAs that had 774 an associated Hfq peak in our RIP-seq dataset. As input, three homologous sRNA sequences were 775 uploaded for each sRNA including the C. difficile 630 homologue. Due to the low sequence 776 conservation, only homologues present in other C. difficile strains could be selected as input 777 sequences, however, we focused on those sequences that showed at least some sequence variation 778 when compared to C. difficile 630. The target region was specified as 200 nt upstream and 100 nt 779 downstream of annotated start codons. Default settings were used. CopraRNA predictions for 780 each sRNA candidate were then compared with mRNA candidates that co-immunoprecipitated 781 with Hfq to filter for high probability targets.

782 DATA AVAILABILITY

783 All RNA-sequencing data are available at the NCBI GEO database 784 (http://www.ncbi.nlm.nih.gov/geo) under the accession number GSE155167.

785 ACKNOWLEDGEMENT

786 We thank the Core Unit SysMed at the University of Würzburg for technical support with RIP- 787 seq and RNAtag-Seq data generation. This work was supported by the IZKF at the University of 788 Würzburg (project Z-6). We thank the Vogel Stiftung Dr. Eckernkamp for supporting F.P. with a 789 Dr. Eckernkamp Fellowship.

790 AUTHOR CONTRIBUTIONS

24 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

791 MF, VL-S, JV, and FF conceived and designed the study. MF and V-LS performed experiments, LJ, 792 V-LS and LB setup the web browser. VL-S, FP and LB analyzed RNA-seq data. FF supervised the 793 project. MF, VL-S and FF wrote the manuscript.

794 REFERENCES

795 Adams, P.P., and Storz, G. (2020). Prevalence of small base-pairing RNAs derived from diverse 796 genomic loci. Biochim Biophys Acta Gene Regul Mech 1863, 194524.

797 Agafonov, D.E., and Spirin, A.S. (2004). The ribosome-associated inhibitor A reduces 798 translation errors. Biochemical and biophysical research communications 320, 354-358.

799 Anjuwon-Foster, B.R., and Tamayo, R. (2017). A genetic switch controls the production of 800 flagella and toxins in Clostridium difficile. PLoS Genet 13, e1006701.

801 Arrecubieta, C., Lee, M.H., Macey, A., Foster, T.J., and Lowy, F.D. (2007). SdrF, a 802 Staphylococcus epidermidis surface protein, binds type I collagen. J Biol Chem 282, 18767- 803 18776.

804 Babina, A.M., Parker, D.J., Li, G.W., and Meyer, M.M. (2018). Fitness advantages conferred 805 by the L20-interacting RNA cis-regulator of ribosomal protein synthesis in Bacillus subtilis. 806 Rna 24, 1133-1143.

807 Bailey, T.L., and Elkan, C. (1994). Fitting a mixture model by expectation maximization to 808 discover motifs in biopolymers. Proceedings. International Conference on Intelligent Systems 809 for Molecular Biology 2, 28-36.

810 Bailey, T.L., and Machanick, P. (2012). Inferring direct DNA binding from ChIP-seq. Nucleic 811 Acids Res 40, e128.

812 Berges, M., Michel, A.M., Lassek, C., Nuss, A.M., Beckstette, M., Dersch, P., Riedel, K., 813 Sievers, S., Becher, D., Otto, A., et al. (2018). Iron Regulation in Clostridioides difficile. 814 Frontiers in microbiology 9, 3183.

815 Bohn, C., Rigoulay, C., Chabelskaya, S., Sharma, C.M., Marchais, A., Skorski, P., Borezee- 816 Durant, E., Barbet, R., Jacquet, E., Jacq, A., et al. (2010). Experimental discovery of small 817 RNAs in Staphylococcus aureus reveals a riboregulator of central metabolism. Nucleic Acids 818 Res 38, 6620-6636.

819 Bordeleau, E., Fortier, L.C., Malouin, F., and Burrus, V. (2011). c-di-GMP turn-over in 820 Clostridium difficile is controlled by a plethora of diguanylate cyclases and phosphodiesterases. 821 PLoS Genet 7, e1002039.

822 Boudry, P., Gracia, C., Monot, M., Caillet, J., Saujet, L., Hajnsdorf, E., Dupuy, B., Martin- 823 Verstraete, I., and Soutourina, O. (2014). Pleiotropic role of the RNA chaperone protein Hfq in 824 the human pathogen Clostridium difficile. J Bacteriol 196, 3234-3248.

825 Bouillaut, L., Dubois, T., Sonenshein, A.L., and Dupuy, B. (2015). Integration of metabolism 826 and virulence in Clostridium difficile. Res Microbiol 166, 375-383.

25 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

827 Caillet, J., Gracia, C., Fontaine, F., and Hajnsdorf, E. (2014). Clostridium difficile Hfq can 828 replace Escherichia coli Hfq for most of its function. RNA 20, 1567-1578.

829 Chen, Y., Indurthi, D.C., Jones, S.W., and Papoutsakis, E.T. (2011). Small RNAs in the genus 830 Clostridium. MBio 2, e00340-00310.

831 Choonee, N., Even, S., Zig, L., and Putzer, H. (2007). Ribosomal protein L20 controls 832 expression of the Bacillus subtilis infC operon via a transcription attenuation mechanism. 833 Nucleic Acids Res 35, 1578-1588.

834 Dambach, M., Irnov, I., and Winkler, W.C. (2013). Association of RNAs with Bacillus subtilis 835 Hfq. PLoS One 8, e55156.

836 Dar, D., Shamir, M., Mellin, J.R., Koutero, M., Stern-Ginossar, N., Cossart, P., and Sorek, R. 837 (2016). Term-seq reveals abundant ribo-regulation of antibiotics resistance in bacteria. Science 838 352, aad9822.

839 Darty, K., Denise, A., and Ponty, Y. (2009). VARNA: Interactive drawing and editing of the 840 RNA secondary structure. Bioinformatics (Oxford, England) 25, 1974-1975.

841 De Mets, F., Van Melderen, L., and Gottesman, S. (2019). Regulation of acetate metabolism 842 and coordination with the TCA cycle via a processed small RNA. Proc Natl Acad Sci U S A 843 116, 1043-1052.

844 DebRoy, S., Gebbie, M., Ramesh, A., Goodson, J.R., Cruz, M.R., van Hoof, A., Winkler, W.C., 845 and Garsin, D.A. (2014). Riboswitches. A riboswitch-containing sRNA controls gene 846 expression by sequestration of a response regulator. Science 345, 937-940.

847 Donnelly, M.L., Fimlaid, K.A., and Shen, A. (2016). Characterization of Clostridium difficile 848 Spores Lacking Either SpoVAC or Dipicolinic Acid Synthetase. J Bacteriol 198, 1694-1707.

849 Edwards, A.N., Nawrocki, K.L., and McBride, S.M. (2014). Conserved oligopeptide permeases 850 modulate sporulation initiation in Clostridium difficile. Infect Immun 82, 4276-4291.

851 El Meouche, I., Peltier, J., Monot, M., Soutourina, O., Pestel-Caron, M., Dupuy, B., and Pons, 852 J.L. (2013). Characterization of the SigD regulon of C. difficile and its positive control of toxin 853 production through the regulation of tcdR. PLoS One 8, e83748.

854 Emerson, J.E., Reynolds, C.B., Fagan, R.P., Shaw, H.A., Goulding, D., and Fairweather, N.F. 855 (2009). A novel genetic switch controls phase variable expression of CwpV, a Clostridium 856 difficile cell wall protein. Mol Microbiol 74, 541-556.

857 Fagan, R.P., and Fairweather, N.F. (2011). Clostridium difficile has two parallel and essential 858 Sec secretion systems. J Biol Chem 286, 27483-27493.

859 Förstner, K.U., Vogel, J., and Sharma, C.M. (2014). READemption-a tool for the 860 computational analysis of deep-sequencing-based transcriptome data. Bioinformatics (Oxford, 861 England) 30, 3421-3423.

862 Frohlich, K.S., Haneke, K., Papenfort, K., and Vogel, J. (2016). The target spectrum of SdsR 863 small RNA in Salmonella. Nucleic Acids Res 44, 10406-10422.

26 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

864 Fujita, Y. (2009). Carbon catabolite control of the metabolic network in Bacillus subtilis. 865 Bioscience, biotechnology, and biochemistry 73, 245-259.

866 Gaballa, A., Antelmann, H., Aguilar, C., Khakh, S.K., Song, K.B., Smaldone, G.T., and 867 Helmann, J.D. (2008). The Bacillus subtilis iron-sparing response is mediated by a Fur- 868 regulated small RNA and three small, basic proteins. Proc Natl Acad Sci U S A 105, 11927- 869 11932.

870 Garai, P., and Blanc-Potard, A. (2020). Uncovering small membrane proteins in pathogenic 871 bacteria: Regulatory functions and therapeutic potential. Mol Microbiol.

872 Gimpel, M., Preis, H., Barth, E., Gramzow, L., and Brantl, S. (2012). SR1--a small RNA with 873 two remarkably conserved functions. Nucleic Acids Res 40, 11659-11672.

874 Hammerle, H., Amman, F., Vecerek, B., Stulke, J., Hofacker, I., and Blasi, U. (2014). Impact 875 of Hfq on the Bacillus subtilis transcriptome. PLoS One 9, e98661.

876 Hemm, M.R., Weaver, J., and Storz, G. (2020). Escherichia coli Small Proteome. EcoSal Plus 877 9.

878 Holmqvist, E., and Vogel, J. (2018). RNA-binding proteins in bacteria. Nat Rev Microbiol 16, 879 601-615.

880 Hor, J., Garriss, G., Di Giorgio, S., Hack, L.M., Vanselow, J.T., Forstner, K.U., Schlosser, A., 881 Henriques-Normark, B., and Vogel, J. (2020). Grad-seq in a Gram-positive bacterium reveals 882 exonucleolytic sRNA activation in competence control. The EMBO journal 39, e103852.

883 Hor, J., Gorski, S.A., and Vogel, J. (2018). Bacterial RNA Biology on a Genome Scale. Mol 884 Cell 70, 785-799.

885 Hör, J., Matera, G., Vogel, J., Gottesman, S., and Storz, G. (2020). Trans-Acting Small RNAs 886 and Their Effects on Gene Expression in Escherichia coli and Salmonella enterica. EcoSal Plus 887 9.

888 Hoyos, M., Huber, M., Förstner, K.U., and Papenfort, K. (2020). Gene autoregulation by 3' 889 UTR-derived bacterial small RNAs. eLife 9.

890 Ingham, C.J., Dennis, J., and Furneaux, P.A. (1999). Autogenous regulation of transcription 891 termination factor Rho and the requirement for Nus factors in Bacillus subtilis. Mol Microbiol 892 31, 651-663.

893 Irnov, I., Sharma, C.M., Vogel, J., and Winkler, W.C. (2010). Identification of regulatory RNAs 894 in Bacillus subtilis. Nucleic Acids Res 38, 6637-6651.

895 J. Sambrook, E.F.F., and T. Maniatis. (1989). Molecular cloning: A laboratory manual. Second 896 edition. Volumes 1, 2, and 3. Current protocols in molecular biology. , Vol 61 (New York: 897 Greene Publishing Associates and John Wiley & Sons.).

898 Jose, B.R., Gardner, P.P., and Barquist, L. (2019). Transcriptional noise and exaptation as 899 sources for bacterial sRNAs. Biochemical Society transactions.

900 Ju, X., Li, D., and Liu, S. (2019). Full-length RNA profiling reveals pervasive bidirectional 901 transcription terminators in bacteria. Nature microbiology 4, 1907-1918.

27 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

902 Kalvari, I., Nawrocki, E.P., Argasinska, J., Quinones-Olvera, N., Finn, R.D., Bateman, A., and 903 Petrov, A.I. (2018). Non-Coding RNA Analysis Using the Rfam Database. Current protocols 904 in bioinformatics 62, e51.

905 Kavita, K., de Mets, F., and Gottesman, S. (2018). New aspects of RNA-based regulation by 906 Hfq and its partner sRNAs. Curr Opin Microbiol 42, 53-61.

907 Khanna, S., and Gerding, D.N. (2019). Current and future trends in clostridioides (clostridium) 908 difficile infection management. Anaerobe.

909 Kint, N., Alves Feliciano, C., Hamiot, A., Denic, M., Dupuy, B., and Martin-Verstraete, I. 910 (2019). The sigma(B) signalling activation pathway in the enteropathogen Clostridioides 911 difficile. Environ Microbiol 21, 2852-2870.

912 Kint, N., Janoir, C., Monot, M., Hoys, S., Soutourina, O., Dupuy, B., and Martin-Verstraete, I. 913 (2017). The alternative sigma factor sigma(B) plays a crucial role in adaptive strategies of 914 Clostridium difficile during gut infection. Environ Microbiol 19, 1933-1958.

915 Kirk, J.A., and Fagan, R.P. (2016). Heat shock increases conjugation efficiency in Clostridium 916 difficile. Anaerobe 42, 1-5.

917 Kroger, C., Colgan, A., Srikumar, S., Handler, K., Sivasankaran, S.K., Hammarlof, D.L., 918 Canals, R., Grissom, J.E., Conway, T., Hokamp, K., et al. (2013). An infection-relevant 919 transcriptomic compendium for Salmonella enterica Serovar Typhimurium. Cell Host Microbe 920 14, 683-695.

921 Kröger, C., MacKenzie, K.D., Alshabib, E.Y., Kirzinger, M.W.B., Suchan, D.M., Chao, T.C., 922 Akulova, V., Miranda-CasoLuengo, A.A., Monzon, V.A., Conway, T., et al. (2018). The 923 primary transcriptome, small RNAs and regulation of antimicrobial resistance in Acinetobacter 924 baumannii ATCC 17978. Nucleic Acids Res 46, 9684-9698.

925 Lawson, P.A., Citron, D.M., Tyrrell, K.L., and Finegold, S.M. (2016). Reclassification of 926 Clostridium difficile as Clostridioides difficile (Hall and O'Toole 1935) Prevot 1938. Anaerobe 927 40, 95-99.

928 Lloréns-Rico, V., Cano, J., Kamminga, T., Gil, R., Latorre, A., Chen, W.H., Bork, P., Glass, 929 J.I., Serrano, L., and Lluch-Senar, M. (2016). Bacterial antisense RNAs are mainly the product 930 of transcriptional noise. Sci Adv 2, e1501363.

931 Lorenz, R., Bernhart, S.H., Höner Zu Siederdissen, C., Tafer, H., Flamm, C., Stadler, P.F., and 932 Hofacker, I.L. (2011). ViennaRNA Package 2.0. Algorithms for molecular biology : AMB 6, 933 26.

934 Mader, U., Nicolas, P., Depke, M., Pane-Farre, J., Debarbouille, M., van der Kooi-Pol, M.M., 935 Guerin, C., Derozier, S., Hiron, A., Jarmer, H., et al. (2016). Staphylococcus aureus 936 Transcriptome Architecture: From Laboratory to Infection-Mimicking Conditions. PLoS Genet 937 12, e1005962.

938 Maikova, A., Kreis, V., Boutserin, A., Severinov, K., and Soutourina, O. (2019). Using an 939 Endogenous CRISPR-Cas System for Genome Editing in the Human Pathogen Clostridium 940 difficile. Appl Environ Microbiol 85.

28 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

941 Maikova, A., Peltier, J., Boudry, P., Hajnsdorf, E., Kint, N., Monot, M., Poquet, I., Martin- 942 Verstraete, I., Dupuy, B., and Soutourina, O. (2018). Discovery of new type I toxin-antitoxin 943 systems adjacent to CRISPR arrays in Clostridium difficile. Nucleic Acids Res.

944 Marchais, A., Duperrier, S., Durand, S., Gautheret, D., and Stragier, P. (2011). CsfG, a 945 sporulation-specific, small non-coding RNA highly conserved in endospore formers. RNA Biol 946 8, 358-364.

947 Martin-Verstraete, I., Peltier, J., and Dupuy, B. (2016). The Regulatory Networks That Control 948 Clostridium difficile Toxin Synthesis. Toxins (Basel) 8.

949 Masse, E., and Gottesman, S. (2002). A small RNA regulates the expression of genes involved 950 in iron metabolism in Escherichia coli. Proc Natl Acad Sci U S A 99, 4620-4625.

951 Masse, E., Vanderpool, C.K., and Gottesman, S. (2005). Effect of RyhB small RNA on global 952 iron use in Escherichia coli. J Bacteriol 187, 6962-6971.

953 McKee, R.W., Harvest, C.K., and Tamayo, R. (2018). Cyclic Diguanylate Regulates Virulence 954 Factor Genes via Multiple Riboswitches in Clostridium difficile. mSphere 3.

955 McKee, R.W., Mangalea, M.R., Purcell, E.B., Borchardt, E.K., and Tamayo, R. (2013). The 956 second messenger cyclic Di-GMP regulates Clostridium difficile toxin production by 957 controlling expression of sigD. J Bacteriol 195, 5174-5185.

958 Melamed, S., Adams, P.P., Zhang, A., Zhang, H., and Storz, G. (2020). RNA-RNA 959 Interactomes of ProQ and Hfq Reveal Overlapping and Competing Roles. Mol Cell 77, 411- 960 425.e417.

961 Mellin, J.R., Koutero, M., Dar, D., Nahori, M.A., Sorek, R., and Cossart, P. (2014). 962 Riboswitches. Sequestration of a two-component response regulator by a riboswitch-regulated 963 noncoding RNA. Science 345, 940-943.

964 Miyakoshi, M., Matera, G., Maki, K., Sone, Y., and Vogel, J. (2019). Functional expansion of 965 a TCA cycle operon mRNA by a 3' end-derived small RNA. Nucleic Acids Res 47, 2075-2088.

966 Nelson, J.D., Kusmiesz, H., Jackson, L.H., and Woodman, E. (1980). Treatment of Salmonella 967 gastroenteritis with ampicillin, amoxicillin, or placebo. Pediatrics 65, 1125-1130.

968 Nicolas, P., Mader, U., Dervyn, E., Rochat, T., Leduc, A., Pigeonneau, N., Bidnenko, E., 969 Marchadier, E., Hoebeke, M., Aymerich, S., et al. (2012). Condition-dependent transcriptome 970 reveals high-level regulatory architecture in Bacillus subtilis. Science 335, 1103-1106.

971 Otto, C., Stadler, P.F., and Hoffmann, S. (2014). Lacking alignments? The next-generation 972 sequencing mapper segemehl revisited. Bioinformatics (Oxford, England) 30, 1837-1843.

973 Peltier, J., Shaw, H.A., Couchman, E.C., Dawson, L.F., Yu, L., Choudhary, J.S., Kaever, V., 974 Wren, B.W., and Fairweather, N.F. (2015). Cyclic diGMP regulates production of sortase 975 substrates of Clostridium difficile and their surface exposure through ZmpI protease-mediated 976 cleavage. J Biol Chem 290, 24453-24469.

977 Peng, Z., Jin, D., Kim, H.B., Stratton, C.W., Wu, B., Tang, Y.W., and Sun, X. (2017). Update 978 on Antimicrobial Resistance in Clostridium difficile: Resistance Mechanisms and 979 Antimicrobial Susceptibility Testing. J Clin Microbiol 55, 1998-2008.

29 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

980 Pettit, L.J., Browne, H.P., Yu, L., Smits, W.K., Fagan, R.P., Barquist, L., Martin, M.J., 981 Goulding, D., Duncan, S.H., Flint, H.J., et al. (2014). Functional genomics reveals that 982 Clostridium difficile Spo0A coordinates sporulation, virulence and metabolism. BMC 983 Genomics 15, 160.

984 Pi, H., and Helmann, J.D. (2017). Sequential induction of Fur-regulated genes in response to 985 iron limitation in Bacillus subtilis. Proc Natl Acad Sci U S A 114, 12785-12790.

986 Pishdadian, K., Fimlaid, K.A., and Shen, A. (2015). SpoIIID-mediated regulation of sigmaK 987 function during Clostridium difficile sporulation. Mol Microbiol 95, 189-208.

988 Pitman, S., and Cho, K.H. (2015). The Mechanisms of Virulence Regulation by Small 989 Noncoding RNAs in Low GC Gram-Positive Pathogens. International journal of molecular 990 sciences 16, 29797-29814.

991 Quirk, P.G., Dunkley, E.A., Jr., Lee, P., and Krulwich, T.A. (1993). Identification of a putative 992 Bacillus subtilis rho gene. J Bacteriol 175, 647-654.

993 Raden, M., Ali, S.M., Alkhnbashi, O.S., Busch, A., Costa, F., Davis, J.A., Eggenhofer, F., 994 Gelhausen, R., Georg, J., Heyne, S., et al. (2018). Freiburg RNA tools: a central online resource 995 for RNA-focused research and teaching. Nucleic Acids Res 46, W25-w29.

996 Robinson, J.T., Thorvaldsdóttir, H., Winckler, W., Guttman, M., Lander, E.S., Getz, G., and 997 Mesirov, J.P. (2011). Integrative genomics viewer. Nat Biotechnol 29, 24-26.

998 Robinson, M.D., McCarthy, D.J., and Smyth, G.K. (2010). edgeR: a Bioconductor package for 999 differential expression analysis of digital gene expression data. Bioinformatics (Oxford, 1000 England) 26, 139-140.

1001 Rochat, T., Bohn, C., Morvan, C., Le Lam, T.N., Razvi, F., Pain, A., Toffano-Nioche, C., 1002 Ponien, P., Jacq, A., Jacquet, E., et al. (2018). The conserved regulatory RNA RsaE down- 1003 regulates the arginine degradation pathway in Staphylococcus aureus. Nucleic Acids Res 46, 1004 8803-8816.

1005 Rochat, T., Delumeau, O., Figueroa-Bossi, N., Noirot, P., Bossi, L., Dervyn, E., and Bouloc, P. 1006 (2015). Tracking the Elusive Function of Bacillus subtilis Hfq. PLoS One 10, e0124977.

1007 Ruiz de los Mozos, I., Vergara-Irigaray, M., Segura, V., Villanueva, M., Bitarte, N., Saramago, 1008 M., Domingues, S., Arraiano, C.M., Fechter, P., Romby, P., et al. (2013). Base pairing 1009 interaction between 5'- and 3'-UTRs controls icaR mRNA translation in Staphylococcus aureus. 1010 PLoS Genet 9, e1004001.

1011 Rupnik, M., Wilcox, M.H., and Gerding, D.N. (2009). Clostridium difficile infection: new 1012 developments in epidemiology and pathogenesis. Nat Rev Microbiol 7, 526-536.

1013 Ryan, D., Jenniches, L., Reichardt, S., Barquist, L., and Westermann, A.J. (2020). A high- 1014 resolution transcriptome map identifies small RNA regulation of metabolism in the gut microbe 1015 Bacteroides thetaiotaomicron. Nature communications 11, 3557.

1016 Saujet, L., Monot, M., Dupuy, B., Soutourina, O., and Martin-Verstraete, I. (2011). The key 1017 sigma factor of transition phase, SigH, controls sporulation, metabolism, and virulence factor 1018 expression in Clostridium difficile. J Bacteriol 193, 3186-3196.

30 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

1019 Saujet, L., Pereira, F.C., Serrano, M., Soutourina, O., Monot, M., Shelyakin, P.V., Gelfand, 1020 M.S., Dupuy, B., Henriques, A.O., and Martin-Verstraete, I. (2013). Genome-wide analysis of 1021 cell type-specific gene transcription during spore formation in Clostridium difficile. PLoS 1022 Genet 9, e1003756.

1023 Schmalisch, M., Maiques, E., Nikolov, L., Camp, A.H., Chevreux, B., Muffler, A., Rodriguez, 1024 S., Perkins, J., and Losick, R. (2010). Small genes under sporulation control in the Bacillus 1025 subtilis genome. J Bacteriol 192, 5402-5412.

1026 Schoenfelder, S.M.K., Lange, C., Prakash, S.A., Marincola, G., Lerch, M.F., Wencker, F.D.R., 1027 Forstner, K.U., Sharma, C.M., and Ziebuhr, W. (2019). The small non-coding RNA RsaE 1028 influences extracellular matrix composition in Staphylococcus epidermidis biofilm 1029 communities. PLoS pathogens 15, e1007618.

1030 Serrano, M., Kint, N., Pereira, F.C., Saujet, L., Boudry, P., Dupuy, B., Henriques, A.O., and 1031 Martin-Verstraete, I. (2016). A Recombination Directionality Factor Controls the Cell Type- 1032 Specific Activation of sigmaK and the Fidelity of Spore Development in Clostridium difficile. 1033 PLoS Genet 12, e1006312.

1034 Sharma, C.M., Hoffmann, S., Darfeuille, F., Reignier, J., Findeiss, S., Sittka, A., Chabas, S., 1035 Reiche, K., Hackermuller, J., Reinhardt, R., et al. (2010). The primary transcriptome of the 1036 major human pathogen Helicobacter pylori. Nature 464, 250-255.

1037 Sharma, C.M., and Vogel, J. (2014). Differential RNA-seq: the approach behind and the 1038 biological insight gained. Curr Opin Microbiol 19, 97-105.

1039 Shelton, A.N., Lyu, X., and Taga, M.E. (2020). Flexible Cobamide Metabolism in 1040 Clostridioides (Clostridium) difficile 630 Deltaerm. J Bacteriol 202.

1041 Shishkin, A.A., Giannoukos, G., Kucukural, A., Ciulla, D., Busby, M., Surka, C., Chen, J., 1042 Bhattacharyya, R.P., Rudy, R.F., Patel, M.M., et al. (2015). Simultaneous generation of many 1043 RNA-seq libraries in a single reaction. Nature methods 12, 323-325.

1044 Silvaggi, J.M., Perkins, J.B., and Losick, R. (2006). Genes for small, noncoding RNAs under 1045 sporulation control in Bacillus subtilis. J Bacteriol 188, 532-541.

1046 Smirnov, A., Forstner, K.U., Holmqvist, E., Otto, A., Gunster, R., Becher, D., Reinhardt, R., 1047 and Vogel, J. (2016). Grad-seq guides the discovery of ProQ as a major small RNA-binding 1048 protein. Proc Natl Acad Sci U S A 113, 11591-11596.

1049 Soutourina, O. (2019). Type I Toxin-Antitoxin Systems in Clostridia. Toxins (Basel) 11.

1050 Soutourina, O.A., Monot, M., Boudry, P., Saujet, L., Pichon, C., Sismeiro, O., Semenova, E., 1051 Severinov, K., Le Bouguenec, C., Coppee, J.Y., et al. (2013). Genome-wide identification of 1052 regulatory RNAs in the human pathogen Clostridium difficile. PLoS Genet 9, e1003493.

1053 Stabler, R.A., He, M., Dawson, L., Martin, M., Valiente, E., Corton, C., Lawley, T.D., Sebaihia, 1054 M., Quail, M.A., Rose, G., et al. (2009). Comparative genome and phenotypic analysis of 1055 Clostridium difficile 027 strains provides insight into the evolution of a hypervirulent 1056 bacterium. Genome Biol 10, R102.

1057 Storz, G., Wolf, Y.I., and Ramamurthi, K.S. (2014). Small proteins can no longer be ignored. 1058 Annual review of biochemistry 83, 753-777.

31 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

1059 van Schaik, W., and Abee, T. (2005). The role of sigmaB in the stress response of Gram-positive 1060 bacteria -- targets for food preservation and safety. Current opinion in biotechnology 16, 218- 1061 224.

1062 Wade, J.T., and Grainger, D.C. (2014). Pervasive transcription: illuminating the dark matter of 1063 bacterial transcriptomes. Nat Rev Microbiol 12, 647-653.

1064 Warrier, I., Ram-Mohan, N., Zhu, Z., Hazery, A., Echlin, H., Rosch, J., Meyer, M.M., and van 1065 Opijnen, T. (2018). The Transcriptional landscape of Streptococcus pneumoniae TIGR4 reveals 1066 a complex operon architecture and abundant riboregulation critical for growth and virulence. 1067 PLoS pathogens 14, e1007461.

1068 Weinberg, Z., Lunse, C.E., Corbino, K.A., Ames, T.D., Nelson, J.W., Roth, A., Perkins, K.R., 1069 Sherlock, M.E., and Breaker, R.R. (2017). Detection of 224 candidate structured RNAs by 1070 comparative analysis of specific subsets of intergenic regions. Nucleic Acids Res 45, 10811- 1071 10823.

1072 Woods, E.C., Edwards, A.N., Childress, K.O., Jones, J.B., and McBride, S.M. (2018). The C. 1073 difficile clnRAB operon initiates adaptations to the host environment in response to LL-37. 1074 PLoS pathogens 14, e1007153.

1075 Wright, P.R., Georg, J., Mann, M., Sorescu, D.A., Richter, A.S., Lott, S., Kleinkauf, R., Hess, 1076 W.R., and Backofen, R. (2014). CopraRNA and IntaRNA: predicting small RNA targets, 1077 networks and interaction domains. Nucleic Acids Res 42, W119-123.

1078 Wright, P.R., Richter, A.S., Papenfort, K., Mann, M., Vogel, J., Hess, W.R., Backofen, R., and 1079 Georg, J. (2013). Comparative genomics boosts target prediction for bacterial small RNAs. 1080 Proc Natl Acad Sci U S A 110, E3487-3496.

1081 Wurtzel, O., Sesto, N., Mellin, J.R., Karunker, I., Edelheit, S., Becavin, C., Archambaud, C., 1082 Cossart, P., and Sorek, R. (2012). Comparative transcriptomics of pathogenic and non- 1083 pathogenic Listeria species. Molecular systems biology 8, 583.

1084 Wydau-Dematteis, S., El Meouche, I., Courtin, P., Hamiot, A., Lai-Kuen, R., Saubamea, B., 1085 Fenaille, F., Butel, M.J., Pons, J.L., Dupuy, B., et al. (2018). Cwp19 Is a Novel Lytic 1086 Transglycosylase Involved in Stationary-Phase Autolysis Resulting in Toxin Release in 1087 Clostridium difficile. MBio 9.

1088 Yu, S.H., Vogel, J., and Forstner, K.U. (2018). ANNOgesic: a Swiss army knife for the RNA- 1089 seq based annotation of bacterial/archaeal genomes. GigaScience 7.

1090 Zheng, J.J., Perez, A.J., Tsui, H.T., Massidda, O., and Winkler, M.E. (2017). Absence of the 1091 KhpA and KhpB (JAG/EloR) RNA-binding proteins suppresses the requirement for PBP2b by 1092 overproduction of FtsA in Streptococcus pneumoniae D39. Mol Microbiol 106, 793-814. 1093

1094 FIGURES LEGENDS

1095 Figure 1 Global RNA-seq approaches for high-resolution transcriptome mapping. (A) Sequencing 1096 samples were generated from strain C. difficile 630 grown to late-exponential and stationary

32 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

1097 phase of growth in Tryptone-yeast broth, as well as late-exponential phase in Brain-heart-infusion 1098 broth. (B) Distribution of newly annotated features (TSSs, TTSs and non-coding RNAs) across the 1099 genome of C. difficile strain 630. (C) Read profiles generated with dRNA-seq and RNAtag-Seq allow 1100 the annotation of transcriptional start sites (TSSs) and transcriptional termination sites (TTSs). 1101 For dRNA-seq one fraction of total RNA is treated with terminator exonuclease (TEX+), which 1102 specifically degr

1103 remains untreated.ades This processed differential transcripts treatment carrying results ain 5′a relative monophosphate. read enrichment The other for primary fraction 1104 transcripts in the TEX+ treated libraries allowing the identification of TSSs. For RNAtag-Seq 1105 adapters for sequencing are ligated to RNA 3’ ends thereby capturing 3’ ends of transcripts. (D) 1106 Benchmarking of dRNAseq approach. Two experimentally determined growth-phase dependent 1107 TSS for spo0A are consistent with dRNA-seq based identification of spo0A associated TSSs. 1108 Abbreviations: LE, late-exponential growth phase; ST, stationary growth phase; CDS, coding 1109 sequence; TSS, transcriptional start site; TTS, transcript termination site.

1110

1111 Figure 2 Features associated with transcriptional start sites. (A) Top: TSS classification based on 1112 expression strength and genomic location: primary (P), secondary (S), internal (I), antisense (A) 1113 and orphan (O). Bottom: Venn diagram showing distribution of TSSs among classes. (B) Re- 1114 annotation of CDIF630_02227 gene based on a newly identified internal TSS 30 bases 1115 downstream of its annotated start codon. Reads of dRNA-seq libraries (TEX+/TEX-) mapping to 1116 the CDIF630_02227 region are shown above the sequence covering the region around the old and 1117 new start codon. The old and new AUGs and associated Shine-Dalgarno (SD) sequences are 1118 indicated. (C) Promoter regions associated with detected TSSs were analysed using the MEME 1119 suite. Promoter motifs were recovered for SigA, SigH, SigB and SigK. (D) Two alternative promoter 1120 sequences for SigA and SigB are associated with a pTSS and sTSS of recA, respectively.

1121

1122 Figure 3 Features associated with transcriptional termination sites. (A) Top: TTS classification 1123 based on genomic location: 3UTR (downstream of CDS or non-coding gene), 5UTR (within the 1124 5’ UTR of a coding sequence), CDS (within coding sequence), orphan (downstream of orphan TSS), 1125 CRISPR (associated with CRISPR array). (B) Frequencies of 3’ UTR lengths based on 1741 TTSs. 1126 (C) Consensus motif associated with 3’ UTR termination sites was identified using the MEME suite. 1127 (D) Reads in the RNAtag-Seq libraries reveal overlapping transcription termination sites for 1128 convergently transcribed gene pairs.

1129

1130 Figure 4 Discovery of cis-regulatory elements in 5’ UTR regions. (A) Frequencies of 5’ UTR lengths 1131 based on 1646 primary and secondary TSSs. Red bars indicate six leaderless mRNAs with a 5’ UTR

33 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

1132 lengths of <10 nt. The inset shows the SD sequence motif of C. difficile 630. (B) Classification of 1133 premature transcription termination (PTT) events in 5’ UTRs of mRNAs. The majority is 1134 associated with known RNA families. The remaining 19 PTT events lack homology to known 1135 riboregulators and are putative novel regulators. (C) Putative L20 leader in the 5’ UTR of the infC 1136 operon. Left: schematic overview of the infC operon organization including 5’ UTR region and 1137 expression analysis of CDIF630nc_00018 by Northern blot. Right: Predicted secondary structure 1138 for CDIF630nc_00018 using RNAfold. Nucleotides are colored according to base-pairing 1139 probabilities. (D) Expression analysis of novel 5’ UTR cis-regulatory elements. Total RNA was 1140 extracted at mid-exponential, late-exponential and stationary growth phases from C. difficile 630 1141 grown in TY medium and analyzed by northern blot using radioactively labeled DNA probes.

1142

1143 Figure 5 The sRNA landscape of C. difficile 630. (A) Classification of annotated sRNA candidates 1144 based on their genomic location. Numbers indicate amount of annotated sRNA candidates for each 1145 class (B) Expression profiles of representative candidates for each sRNA class. Total RNA was 1146 extracted at mid-exponential, late-exponential and stationary growth phases from C. difficile 630 1147 grown in TY, TYG and BHI medium and analyzed by northern blot using radioactively labeled DNA 1148 probes.

1149

1150 Figure 6 The spectrum of Hfq-associated RNA classes in C. difficile 630. (A) Hfq RIP-seq was 1151 performed on WT and 3xFLAG tagged hfq strains grown to mid-exponential, late-exponential and 1152 early-stationary phases in TY medium in two independent experiments. (B) Overview of RIP-seq 1153 workflow. (C) Pull-down of 3xFLAG tagged Hfq was validated by Western blot using anti-FLAG 1154 antibody. Hfq-associated RNAs were validated by northern blot using specific DNA probes. (D) Pie 1155 chart for Hfq RIP-seq showing the relative amount of Hfq-associated sequences mapping to 1156 different RNA classes. (E) Bar chart showing the fraction of sRNAs for each class that were bound 1157 by Hfq in vivo. (F) Number of Hfq-bound sRNA at different growth phases. Abbreviations: ME, mid- 1158 exponential; LE, late-exponential; ST, stationary; IP, immunoprecipitation; L: lysate; E: elution.

1159

1160 Figure 7 Prediction of the target spectrum for the intergenic sRNA AtcS based on Hfq-associated 1161 mRNAs. (A) Number of Hfq-bound mRNA at different growth phases. (B) Scatter-plot analysis of 1162 RIP-

1163 Benjaminiseq results-Hochberg for mRNAs corrected that P were-value enriched ≤0.1) in inthe stationary 3xFLAG tagged phase Hfq(log2 samples. f.c. ≥2; cDNAGenes readencoding ≥10; 1164 for sporulation-associated proteins are labelled. (C) Predicted secondary structure for AtcS using 1165 RNAfold. Nucleotides are colored according to base-pairing probabilities. Nucleotides predicted 1166 to interact with target mRNAS (seed region) are located in an accessible single-stranded stretch

34 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

1167 at the 3’ end (red box). (D) Expression profiling of AtcS by northern blot. Total RNA was extracted 1168 at mid-exponential, late-exponential and stationary growth phases from C. difficile 630 grown in 1169 TY, TYG and BHI medium and analyzed by northern blot using radioactively labeled DNA probes. 1170 (E) Predicted RNA duplexes formed by AtcS with selected mRNA targets. Red letters indicate the 1171 start codon. (F) MEME-based definition of a consensus motif in predicted binding sites of AtcS 1172 mRNA targets. Abbreviations: ME, mid-exponential; LE, late-exponential; ST, stationary.

35 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in Figure 1 perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

A 1.2 LE ST D 1.0 0.8 10 BHI late-exponential Tex(-) 8 0.6 6 4 RPM 2

600 0.4 0

OD BHI TY 400 BHI late-exponential Tex(+)

0.2 TYG 300 RPM 200 100 0 4 8 12 16 20 24 0 time [h] 10 TY stationary Tex(-) 8 6 4 B 0 RPM 4 Mb 2 0 0.5 Mb 80 TY stationary Tex(+) 60 3.5 Mb CDS (+) 40 CDS (-) RPM 20 TSS (+) 0 CP010905.2 1 Mb TSS (-) 4274782 bp 15 TY late-exponential Tex(-) TTS (+) 10 TTS (-)

3 Mb RPM 5 ncRNA 0 1.5 Mb 200 TY late-exponential Tex(+) 2.5 Mb 150

2 Mb RPM 100 50 0

2500 TY late-exponential RNAtag-seq C transcriptional start site 2000 processing site 1500 1000 3'end RPM 500 3'end 0 pTSS sTSS

Relative read depth Relative read depth spo0A

CDS 16S rRNA

SigH TCTAGGAGGAATATAATTTTGGAGTGTCGAATATGCTTTAG Tex (+), 5'PPP-enriched RNAtag-seq SigA GATATGGTATTTTTATAGATGAAATGATAAAATTGTAGGTG Tex (-), no treatment -35 -10 +1 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this Figure 2preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. A C sigA promoter motif (1188/2293) 2

P P P, I O 1

I S bits

ORF 1 ORF 2 ORF 3 0 -35 -10 A A sigB promoter motif (276/2293) 2 Internal (470) Antisense (129) 1 bits Primary (1627) Secondary (187) 318 81 0 -35 -10 126 2 9 sigH promoter motif (69/2293) 1455 00 154 2

0 1 46 24 bits 0 0 Orphan 0 -35 -10 0 78 sigK promoter motif (51/2293) 2 total TSS = 2293 1 bits

0 -35 -10

B D 0 4 BHI stationary Tex(-) -5 3 -10 2

-15 RPM RPM -20 1 -25 BHI stationary Tex(-) 0 0 50 BHI stationary Tex(+) -50 40 -100 30 -150

RPM 20 RPM -200 10 -250 BHI stationary Tex(+) 0 0 1 TY stationary Tex(-) -1 0.8 -2 0.6

RPM 0.4 RPM -3 0.2 -4 TY stationary Tex(-) 0 0 15 TY stationary Tex(+) -20 10

-40 RPM RPM -60 5 -80 TY stationary Tex(+) 0 0 4 TY logarithmic Tex(-) -5 3

-10 2 RPM RPM -15 1 -20 TY logarithmic Tex(-) 0 0 60 TY logarithmic Tex(+) -50 50 40

-100 30 RPM

RPM 20 -150 10 -200 TY logarithmic Tex(+) 0

iTSS pTSS sTSS

CDIF630_02227 recA

SigA AAATTGACTAGGTTTGATTTATGTTATATAATTAATATA TTGTGATTCTTAATGATACAATATTACTATACAAAAAAAGAATGGGGCGTAGTTATGGAG SigB TAAAATTTTAAAAATGTAAAAAAGGGTATACTATTATAG OLD SD OLD Start iTSS NEW SD NEW Start -35 -10 +1 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this Figure 3 preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

A 3UTR B 1741 consensus motif 15 bp downstream of 3UTR termination sites (1372/1710)

2

CDS 1 bits 5UTR 181 CRISPR Orphan 88 13 19 ORF 0 1 2 43 5 6 7 14 151312111098 TSS

Total termination sites: 2042

C D

400 RNAtag-seq 300 400 200 8 > 500 nt 100 0

6 RPM -1 300 -2 4 -3 -4 2

number of mRNAs toxA tcdC 200 0 20 RNAtag-seq 600 800 1000 1200 1400 15 length of 3' UTR 10 number of mRNAs 2 100

0 RPM -50

-100

0 -150 0 200 400 600 800 1000 1200 1400 length of 3' UTR CDIF630_03029 cwp19 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in Figure 4 perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

A B

250 cis-regulatory elements c-di-GMP-I, 11 consensus Shine-Dalgarno conserved 2 novel c-di-GMP-II, 4 cobalamin, 3 200 FMN, 3 glmS, 1 1

Bits glycine, 3 L10 leader, 1 A G lysine, 4 AT T A GTA G TA G TATAT 0 G pan, 1 150 G 12 3 G4 5 6 7 PreQ1, 3 purine, 4 leaderless mRNA 18 79 SAM, 6 100 SECIS4, 2

number of mRNAs speF, 1

50 T-box, 25

0 TPP, 3 0 100 200 300 400 500 600 yjdF, 1 ykkC-yxkD, 2 5' UTR length yybP-ykoY, 1

120 A C G G G G base pairing C U C U G A probability C G 100 A U C U A 0 1 C G nc018 - L20 leader U C 110 U A G U C G A U C G 130 A C C A U A G U A A A U A 90 A G infC rpmI rplT U A A A A A G 140 C

A U A A A U A 150 G C A U U A A U A U A U U U U U A G U G A A BHI TY A U TYG U A A A 160 A G U A G 60 U U U U 80 U U A [nt] G A ME LE ME LE A A ST ST ME LE ST A A A A 50 U A G U U G G A A A U

70 A U 10 U A A U 147 nc018 U A C G A A 40 A U G U U U U U U U G U 170 G A A U A A A U U A G C 20 A U G C A U

A U A A A C A A A A A A U G G U 1 30 179 D TY TY TY TY [nt] ME LE ST [nt] ME LE ST [nt] ME LE ST [nt] ME LE ST 1180 190 242 1180 692 speF RSW 692 501 190 501 404 147 404 331 331 TY 242 nc084 111 [nt] ME LE ST 242 nc008 147 190 190 nc010 147 147 111 67 nc086 111 111 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted AugustBHI 10, 2020. TheTY copyright holderTYG for this Figure 5 preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND[nt] ME 4.0 International LE ST licenseME. LE ST ME LE ST

A nc022

3UTR 67 18 5UTR IGR 3 8 srlB nc021 srlD ORFORF 1 ORF 2 BHI TY TYG [nt] ME LE ST ME LE ST ME LE ST 13 Antisense 242 nc079 B BHI TY TYG [nt] ME LE ST ME LE ST ME LE ST SECIS_4

501 nc079 grdB 404 331 BHI TY TYG [nt] ME LE ST ME LE ST ME LE ST 242

190 nc090 67

147 nc090 secA2 nc087 111 BHI TY TYG [nt] ME LE ST ME LE ST ME LE ST

nc089

67 242

nc089 CDIF630_03016 pfo nc087

BHI TY TYG BHI TY TYG [nt] ME LE ST ME LE ST ME LE ST [nt] ME LE ST ME LE ST ME LE ST nc001 - T2 1180 242 881 nc001 - T1 692 501 190 404 T1 T2 331 CDIF630_00250 nc001 242 nc009 BHI TY TYG 190 [nt] ME LE ST ME LE ST ME LE ST

111 ldhA nc009 nc006

BHI TY TYG [nt] ME LE ST ME LE ST ME LE ST cbiO nc006 190

BHI TY TYG nc021 [nt] 147 ME LE ST ME LE ST ME LE ST

CDIF630_00870 67 nc028 nc021 iscS1

CDIF630_01016 nc028 bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this Figure 6 preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license.

A C Hfq-FLAG Hfq-ctrl mid-exponential late-exponential stationary Hfq-ctrl Hfq-FLAG Hfq-ctrl Hfq-FLAG Hfq-ctrl Hfq-FLAG 1 ME LE ST [kDa] L E L E L E L E L E L E 70 55 600 0.1 35 OD 25

0.01 0 5 10 15 20 15 Hfq time [h] 242 nc079 cis-antisense B 190

lysate 147 cell lysis nc070 3'UTR

RNAprotein 111 IP

67 nc028 3'UTR

3x wash 111 nc006 3'UTR 331

elution nc001 242 3'UTR

190 RNAprotein

D E F Hfq-bound not Hfq-bound ME

3UTR, 159 0 antisense IGR 0 1 CDS, 179 sRNA, 26 16 antitoxin, 13 5UTR 3 4 2 5UTR, 124 tRNA, 3 3UTR ST LE 0 2 4 6 8 10 12 14 16 18 number of sRNA candidates sRNA bioRxiv preprint doi: https://doi.org/10.1101/2020.08.10.244764; this version posted August 10, 2020. The copyright holder for this Figure 7 preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made available under aCC-BY-NC-ND 4.0 International license. A B ME Hfq-associated mRNA during stationary growth 14 CDS 12 3UTR 84 5UTR 10

5 37 8

70 CPM 2 6 18 41 77 log 4 oppB ST LE 2 sigGspoVB spo0A mRNA 0 0 1 2 3 4 5 6 7 8 9 10 11

log2 FC

U A U C C G G 40 G A G G 60 U U A A U U A A G U G 50 G U U U A A U C U C U G G U A U U U 30 A G C A A U U 70 A U U U A U A U A 20 U U A A C U U C C G C A G U C A G U 150 U 80 U C A A A G A A A A U C G U C G C G C A U U 140 A U A A G C G 180 A G U U A A U U U 120 10 A A 110 U A C G G C base pairing C G U A U A U A 190 A U G C U A A U probability U G A U A U seed region G C 0 1 U A 90 A U U A A U A U A A A U A U G C A A A A A A G A C A U U A C C U U A U C U A A C C C C U U U G U U U U U U A U U 1 100 130 160 170 200

D E BHI TY TYG [nt] ME LE ST ME LE ST ME LE ST oppB (CDIF630_00972) 242 -14 -2 AtcS | | 5'-AUA...ACAAA GUUUG...ACC-3' mRNA 190 GGGGGGUUGGG ::||||||:|: UUCCCCAAUCU CDIF630_02859 3'-UUA...AGAGU AAUAA...AUA-5' sRNA | | AtcS trpS 172 160

CDIF630_03429 F -14 -6 target sites | | RBS G GGGGGU UGGG CDIF630_00972 (oppB) 5'-UUA...UAAGG GAAUU...CAG-3' mRNA GGAGGGG -194 U UCGGGU UUA CDIF630_00926 ::||||| RBS GG AGGGG CDIF630_03429 UUUCCCC RBS G AGGGGU AGA CDIF630_01158 3'-UUA...UAGAG AAUCU...AUA-5' sRNA RBS G GAGGGG CDIF630_01691 | | RBS UUA AGGGGG CDIF630_01285 (rnfD) 148 133 -118 A AGGGGU U CDIF630_02255 (argC) RBS G AGGGGG CDIF630_00623 CDIF630_01158 -13 -2 RBS CDIF630_01900 (mogA) AG AGGGGU GAGA | | RBS A AGGGGU GAGAUU CDIF630_03096 5'-UGG...UUUAG AUUUG...UAU-3' mRNA RBS GG AGGGGA AAGA CDIF630_00633 GAGGGGU AGA RBS GGGGGAAGU ACGUGU GAAG CDIF630_01459 :|||||| ||| 2 UUCCCCA UCU 3'-UUA...AGAGU A AAUAA...AUA-5' sRNA 1 | | bits

U A G 183 160 GACG UG 0 1 2G 4 G5 63